Science needs more research software engineers
nature.com
nature.com
When the software industry says to you: "We will nearly triple your salary, you don't have to work weekends, and you also don't have to feed the mice on a Sunday night."
You will 100% take this deal.
I was a 10yr+ academic tool maker in biochemistry, built cutting edge microscopes, hardware, and image analysis software. My lab was successful in our field. I got some papers out of the deal. I also saw things that no human had seen before in a microscope. I worked with very interesting people from around the world. The work in academia is great. You're moving the needle, new data, new failures. These are the perks. It is also highly possible that you have complete creative control of your project. I did, and it was amazing. Custom designed UIs to streamline our experiments, derived from watching students use the system to do their work. A decathalon of software design.
Some reality: Your PI and organization will never compensate you the way the software industry will. In pay, expectations, and benefits. When you're over 30, and you don't technically have a real 401k, you are still paying your student loans that you needed to get into this field, and you're still in the same shitty apartment, something has to give.
Comparison is the thief of joy, and when you see your cohort of computer science graduates your joy will be stolen :). It's good, honest work. A short tour of duty would be useful, and can teach you the difference between splitting the atom and splitting the check.
Academia, at least in bioscience, is still very much an ivory tower. You don't have enough letters after your name to matter, and you will likely be a pet instead of a peer.
Don't stay underwater for too long. Life is short. :D
Our software was a custom C+Win32 app that was ported from CodeWarrior on MacOS 7/8. Windows timers were so crap that I ended up using Ryan Geiss' timer from Milkdrop: http://www.geisswerks.com/ryan/FAQS/timing.html
Yes, that Ryan Geiss, the Winamp one. He now works for Nvidia I think.
Our machines were bristling with serial ports, driving high speed filter wheels, lasers of every color we could get our hands on, special shutters coated with teflon, fast expensive cameras, and more! Their work is very much in my old field, I was in bioenergetics, specifically mitochondria and their dysfunction.
Thanks for the link down memory lane!
Later as I started to hear more about how economies function and how revenues really build up and stem from consumers in volume I came to realize that things that ultimately benefit a lot of people generate a lot of money. I saw an intermediary of this working at Apple and seeing how much funds they had to spend vs biomedical companies that were more conservative with their funds. Consumer electronics and consumer products in general have a lot of customers, benefit a lot of people, and ultimately earn a lot of money.
Academia is much more limited in its scope and immediate benefit. That delay in benefit shapes the money involved in all sorts of surprising ways that aren't immediately apparent while still under the wings of the academic world and the "currencies" they operate with be it notoriety, prizes, grants, etc etc. Ultimately the results and products of academia are suspect and risky since they're often in the prototype unvetted phase of birthing into existence. Those thoughts Elon Musk shares while touring Starbase about design vs manufacturing in the gauntlet of tests against reality where the two forms get vetted side by side come to mind here. His statement of "design is overrated" probably has a close analogue in academia. Products of the mind are essentially untested and may not stand up to reality along whatever dimensions one needs to evaluate them against, or as is probably more often the case simple don't scale to the degree needed to impact a large number of people in a short time frame to translate to paying customers.
Software development is important to more and more industries and the pay disparity caused by insanely large funding and little requirement to produce profits means that other sectors are being priced out of in-house development, especially niche use cases. The ongoing rise of no-code development will be increasingly useful across all sectors but will fail to deliver a lot of these niche applications.
Salary isn't everything. European engineering is a pretty different culture.
The way the us tries to prevent this is by crippling their people with student debt.
This is an underrated point. This is the case for programmers in finance as well, and requires a hefty salary premium to put up with.
Moving away from data collection to analysis, the fraction of PhDs went up, but only reached 1.0 when considering the sub-group specialising in structural biology.
I'm in Europe. My salary is definitely better than the PhD students' salaries, and I have a proper adult pension as that's a legal requirement here. My salary is approximately equal to what a graduate might earn 1-2 years after graduating in the local market, so doesn't match my actual experience, but I accepted the post for pandemic-induced reasons. Certainly the salary does not, nor will ever, compare to levels.fyi/FAANG type jobs or a large corp in country.
However it is true that my position officially is very much a curiosity. We don't have a defined RSE type role, so the slot I fit in is "staying on to help out on project after graduating". My job is a fixed term contract that can only be renewed a certain number of times and I'm approaching that limit soon. There isn't any viable track to joining the ranks of researchers - I would have to do a masters first, and this ironically would require doing an internship, in spite of the fact I have more actual industry (non-university) experience than the entire lab combined.
I'm also not sure if my lab head bent the rules or not on hiring me - it might be the case that I am supposed to have a PhD or at least a masters.
I would agree with top level post in most points. It is interesting work, but I don't "belong" anywhere in "the system". This might change in 10-20 years. Artefact evaluation is very much becoming a thing in systems research, because being able to reproduce other people's work is quite important, and very occasionally you will stumble upon papers whose claims are, ah, more expansive than the associated github project can fulfil. As more research relies on software that graduate students are simply ill-equipped to write (by virtue of having no experience in anything and by being taught by professors most of whom no experience writing production code) the role of an RSE might become more important in time, but like anything it'll be a slow change.
When talking about the "local market" in Europe one needs to take into account the large number of "dark matter devs" that are working remotely for SV companies, at SV salaries. They simply won't ever show up for interviews at local companies.
Pay is quite good, though, so I can't complain.
Admittedly, my path was convoluted; I started as a engineer to help with non-research software at a large lab, and got pulled on to projects via reputation. But I was replacing a Master's student who was essentially at the same academic level as me anyway. It does pay less, but I made the tradeoff for the quality of projects, which was worth more to me at this point in my career. It's still much more than I need, just not at industry levels.
And secondly, while your PHD peers may earn more than you, they also often earn much less than software industry averages.
> At UC, research assistants typically make more than Grad Students or post-docs
That's cool but I didn't say every university; I don't think one counterexample makes my comment "[not] true in the slightest".
Also, imagine all the people who failed out of masters or PhD programs who end up in management and are resentful. It's a surprisingly common thing.
Everyone “scientific programmers” potion I’ve seen wants you to have a PhD and be a domain expert.
I worked on a tech-heavy project large enough to get an NYT feature article covering its launch. For it, I collaborated heavily on the service design and logistics, and singlehandedly designed, built, administered, documented, supported, and provided training for the technical infrastructure and more than a dozen related interfaces and tools. In lines of code, it probably landed somewhere in the low 5 figures, but that was certainly way more than it needed to be. It was hackish but durable and performant. It was an exercise in pure generalism— no individual accomplishment was close to technically innovative enough to warrant a novel white paper, but I was invited to speak at a few related conferences about it.
But the professor overseeing the project didn't even mention me or my role in his launch party speech for the folks in our building, let alone anywhere that would have provided career visibility. He thanked and spoke about the contributions of every other major contributor— even the temp worker who ran the machines (he wouldn't want to appear classist after all)— but I got a hand shake and quiet thank you after his speech for my 5 year effort. I was at every related manager's meeting and largely seen as one of three "go-to" people for the project in general, not just tech stuff.
This sort of gatekeeping is a part of academic culture I just don't get. At least in business there's some predictability to people stepping on each other to get to the top, but what's the purpose of this?
There are so many academic journals, from scammy, to bad (yet honest), to average, to good, to the top. You can publish almost anything, if you select an appropriate, less prestigious journal.
This is just a hypothesis, but I'd predict a high correlation between becoming an academia lifer, and having certain preexisting personality disorders, stemming from having never derived a sense of self worth from anything other than academic achievement since they learned to speak. Or maybe I'm just speaking for myself :)
Similar to the top tier of tech companies being destructive and amoral in their own ways, not only because they're corporations, but also because programmers see technical challenges waiting to be solved like a moth sees a porch light, but see ethical problems dimly. (still probably speaking for myself...)
I'm sure they also didn't think the electricity company for keeping the lights on, or Microsoft for creating Windows to write their speeches, or the guy that emptied the waste baskets in the office so the PHD guy didn't have to.
Don't carry water for someone else. Enrich yourself. That's all anyone else is doing, all the 'research' is for personal enrichment and prestige. Don't prop up the broken academic industry with less than market wages, let them fail.
That's my number one advice regarding academia: unless there's a path toward a valuable visa, or it's paid work while getting a valuable degree (read, something that will have the prestige to open doors) or your co-author at a good university you're much better building something for yourself somewhere else.
> no individual accomplishment was close to technically innovative enough to warrant a novel white paper [...] But the professor overseeing the project didn't even mention me or my role in his launch party speech for the folks in our building, let alone anywhere that would have provided career visibility. He thanked and spoke about the contributions of every other major contributor
That's because papers are the metric by which visibility is measured. Pretty much the only way to move forward is getting your name as author on the main papers.
Anyway, fast forward about 10 years to a few months ago, I get a generic "cold call" type message on LinkedIn from a unicorn data tools company, from the same CEO guy- he bounced around and somehow landed a sales role there. I ignored the first one... he sent a followup, and I was incredulous- did he not remember me? Did he not care? It was something along the lines of "Hey how are you? I am working with xxxx and think you would be interested- can we set up a chat..." and I just replied back saying "I am great, haven't been screamed at or had anyone throw something at me in ten years..." and he still had the balls to right back something like "lol, great. Let me know when we can set up a call..." and I wasn't really sure how to respond, but after about a week just wrote "If your next message isn't a very specific apology for your past behavior, do not contact me. I am surprised with your past attitude you would even work at a place like xxxx." If he was a dick I was going to reach out to their head of sales and possibly CEO explaining his past and how I was disappointed that a firm with their reputation would even hire someone like that. He gave me just a half hearted enough non-specific apology to not do that- followed up immediately with an offering of buying me a beer (so he can pitch me), so I opted not to escalate any further.
I have a few other stories similar to this, where Karma really bit those that were hostile and condescending towards technology and technologists, but this is the most direct. I kept in touch with a few of the old "traders" I used to work with, and used to go out for drinks with them from time to time, and they would invite a larger group of people, and I actually stopped because they were all kind of depressing- they all lost their old jobs, a few pivoted into different decent roles, but mostly they just got drunk pining for the good old days to come back.
While there was a tiny bit of schadenfreude, in the end its just sad.
I have worked for almost 15 years in academic research, but in very close collaboration with the steel industry. The code we write can help steel companies to save millions when developing new products. This is quite complex software, which combines materials science, mechanical engineering and advanced mathematical concepts while requiring high performance and reliability.
I found a nice position for a tenure track in France, in a top research centre. Besides designing and writing software, I would have to design and implement experimental plans, teach, deal with students and administration, keep an excellent publications record, and find funding for future projects. Remote work would not be a possibility (but I would work may unpaid extra hours at home). And the amount of published papers and recommendation letters required just to be considered for the job was overwhelming. My salary would be lower than $30k/year. They do not even know what is a RSE.
I am searching a remote job in the software industry now.
Also, assistant professors (or the equivalent there) generally make less but do probably make more once they get tenure. I'm assuming they meant the tenure-track position itself is ~30K USD, but making tenure usually does mean a pay increase.
30k for a tenure track position sounds insane to me.
That was very lucky, many programs do not offer stipends and require people to take out loans.
Then there are PhDs that get a scholarship (bursaal), that is only around 24000-25000 gross per year.
Not too long ago, there were only employee PhDs, but some universities really love the scholarship system, because they have to pay less tax, so it's a lot cheaper for the universities.
My wife had a PhD scholarship in NL and it really had some large negative effects after finishing her PhD:
- She contributed 4 years less into a pension fund, since bursary PhDs do not build up pension outside the state pension;
- In her next academic position, they didn't consider her four years of PhD work as working experience, while they did do that for me as an employee PhD. So, she was set back 4 years in salary growth.
- She finished her PhD in August and started a full time job after her PhD. Because she had an income that went over some threshold, she had to pay the taxes that the university dodged by using the scholarship system. She worked the rest of the year at a loss (the taxes were higher than the income from September-December).
The worst part of it is that many foreign PhD students do not realize that there is a two-class system.
https://ukrant.nl/minister-zet-definitief-streep-door-experi...
Paid?
I think that's incorrect. You would work the number of hours you wish to work (considering you produce reasonable value, but the bar is low). Research engineer (or researcher for that matter) in a public French research center is a civil servant position. They are difficult to get but you don't get fired unless something is blatantly wrong.
Source: I worked 10+ years in such a position. I work now for a FAANG and the pressure is considerably higher. Evaluations every 6 months, lot of peer pressure (engineers are on average better and more ambitious than those in academia and you need to keep up - some of them seem to work 24/7), extremely stressful oncalls. Gross salary is 5 times my previous salary and has the potential to increase much more.
Of course, this is certainly not representative of all cases, but most of the time, there's a price to pay for a higher salary. Another thing to think about is ageism: as a research engineer in academia, you're all set until retirement. In software industry, it's getting hard after 50.
BINGO! that field is notoriously horrible and interacts extremely badly with a 'when not discovered here, not important' syndrome. Biology is brutal toward bio-physicists, mathematicians as well as people who code who they are forced to work with rather than seek out to help them.
I still hold up as an example nonsense discussions around p-values in bio vs actual work going on in statistics in maths departments. It shows how far detached they've become.
Not to criticize too strongly, but given the above, combined with it's reproducabilty crisis, and existential problem of being in the back-pocket of big-pharma, I seriously doubt the professional integrity of a lot of people in the field.
Move toward mathematics, physics and chemistry. There is (some) serious money and a good atmosphere around areas such as finite-element modelling, or wolphram like tools as an example. There is a lack of direct funding for decent posts but you get recognised and paid the equivalent as a peer, I know from working with some of these people. It's not to say it's 100% always without friction, but no job is I'd argue.
FEM can offer money but you are competing against engineers who that's what they've done for years.
If you interviewed software engineers and data scientists right now I bet a third of them once were physical scientists/mathematicians who mostly regret their degrees or the fact they can't find survivable work using them.
Would mathematicians truly be regretting their degrees, if they decide to work in software? I read that mathematics one of the best degrees for a career in software engineering, as computer science is very closely related to mathematics (to the point where studies of algorithms are largely the same for mathematics and computer science students).
It is also possible that aptitude for math is related to aptitude in software engineering.
However: The mathematics content of 90%+ of mathematics degrees awarded is fully irrelevant to 95%+ of software development tasks. And when that 5% task needs that some kind special mathematical insight, the people who want that task done are going to get the top professional they can find for it. Maybe the prospective math student is going to be that professional, but I don't recommend planning a career for it.
I am not saying there isn't work where some math is useful but the most commonly used applied stuff ... say, linear algebra ... is typically covered in a respectable engineering program; degree in mathematics would be superfluous. Proving theoretical properties of Hilbert spaces or measurable sets or bifurcations of dynamic systems or advances in differentiable topology or fascinating behavior of cellular automata or whatever is going to be gigantic waste of your time if you won't use it later in your career or you don't find it intrinsic motivation in itself.
In my case, I was in basic science which hit a crisis near 2008 when the NIH was flat funded. This caused a come to Jesus moment, where suddenly all basic science labs were rebranded as translational medicine. My department was absolutely gutted, down from 15 or so PIs to maybe 8ish in the span of a year. Our field was bioenergetics which at the time was pretty competitive, and easy to link to diseases/metabolic disorders. We didn't work with pharma, some labs received contracts for small work. NIH was by far the biggest funder, followed by DARPA and other smaller health organizations.
Lack of professional integrity is a very real problem.
Over the past two years I wrote fairly frequently about some of the nonsensical / pseudo-scientific COVID papers that got published, especially the quality problems in epidemiology. Epidemiology isn't bioscience (actually that's one of the problems with it - total lack of biology), but it's adjacent. After that I got contacted by a former research software engineer who worked with top epidemiology teams in the UK. I also got contacted by a member of the SAGE committee.
Both of them told me some absolutely mind-blowing stories of ethical malpractice. I wasn't totally surprised because it was obvious that those sorts of things must have been going on behind the scenes just from reading their model source code, reports, watching their behavior etc. The RSE had become so disgusted at what he'd seen that he actually left the country and switched from working at Oxford to some US university I'd never heard of, switching fields along the way too. Quite the downgrade in prestige but after years of trying to help epidemiologists he concluded the entire field was utterly morally corrupt and he wanted nothing to do with it.
Here are some of the more memorable things I was told by those two scientists:
- The RSE at one point found a bug in a FORTRAN model being used to model malaria outbreaks. It had been used as the basis for hundreds of papers but at critical points was using pointer values as variables instead of dereferencing them. Obviously, a typical pointer has a very different value to most organic things (some FFI bug). He reported this bug to the authors and got a reply back within 30 minutes saying they'd checked the papers (all of them) and it didn't affect the results. This claim was very obviously a lie: not only could they not possibly have checked even one paper in 30 minutes but he already knew fixing the bug did indeed change results! They didn't care and he was shocked that his "colleagues" would bullshit him so directly, especially as they must have known that he would know.
- Same guy flagged code quality issues to some of the scientists and proposed introducing some rules designed to improve quality. He was dismissed with the words: "oh <name>, we're scientists, we don't write bugs".
- The SAGE member told me about some of the internal discussions they had. Criticisms of the methodological validity and accuracies of their models were dismissed with reasoning like this: "that person reads the Spectator so it doesn't matter what they think". Relatedly, he made clear that the supposedly scientific SAGE predictions were sometimes being altered to reduce criticism of the group by left wing media and journalists. The changes were presented as "the science changed" but that wasn't what was going on behind the scenes.
- Malaria research is (supposedly) being badly distorted by the Gates Foundation. Gates only cares about eradication which leads to lots of problems. There are some smaller ones, like many researchers don't genuinely believe that's possible but lie on their grant applications to make mitigation efforts sound like eradication efforts. And then there were unethical experiments on entire populations where e.g. whole areas are blanketed in anti-malarial drugs. If it works, great, you eradicated malaria in that area. If it doesn't you just selected for drug-resistant mosquitos and now the drugs that were being used only to treat the serious cases don't work for anyone. He told me this has actually happened more than once.
- The RSE told me they'd at one point tried to recruit an RSE working with climatologists to help them with their modelling (a belief that climatologists are more rigorous than they are seems to be common in epidemiology). The RSE they interviewed refused to take the job. His reason was he was quitting academia entirely, as he was so disturbed by the practices he'd seen.
A few years ago if you'd told me that a whole research field could be unethical I'd have thought you were crazy because, well, that's a whole lot of people being painted by a very broad brush. Now I've seen it for myself and heard from other former insiders, it's easy to see what happens - the honest ones discover what's happening and leave. Because academia hardly ever penalizes intellectual dishonesty, the pool of people who remain are the ones who are OK with it and have learned that it works / has no consequences. Things steadily become more and more toxic.
What I can and will say (and is on record) is that reproducibility was not a concern from the Imperial College virology dept.
One of the serious downsides of working in academia is you are basically doing the industry's work for them for less pay and they will one day turn around, pat you on the back, then sell your work for millions of dollars. It gets worse honestly the closer you are to applied fields. There, you already straddle the line between what your more "pure" (and less well paid) peers think is "science" and actually making things that will in fact make people's lives better, so you have less room to be idealistic about why you are doing what you're doing, that is, whether it is for "moving the needle" or "adding to the corpus of humanity's knowledge" or whether you really are just doing someone else's work for them they aren't willing to fund given the risks. And given that the latter is basically closer to what you're doing and your closer to the place where you'll see your work enable someone else's riches, it's hard not to want to jump ship and just become one of those people on the other side but make money hands over fist.
It's an upsetting situation honestly.
Perhaps the characterization is wrong and the EV (expected value) of this work is far lower than you think (this seems likely), and/or there are externalities like regulation, or the leverage of prestige that traditional orgs (e.g. universities and publishers) wield, that warp the profit incentive. Or (and this is my cynical view) pure science was always best left to the hobbyists. Historically the most important discoveries have come only rarely and to those who loved doing science in their free time or, more rarely, when a talented individual found a patron. Building a science factory and hiring science factory workers not only sounds distasteful, but it doesn't seem to work very well. (The exceptions being those very capital intensive projects like the LHC which require a large pool of professional scientists and engineers to do the experiment.)
You are the vehicle by which neo-liberal capitalism fixes the problem. By leaving academia to work for a firm directly, you are responding to a price signal sent by the industry, relaying that price signal to the academic labs.
You might object, this is slower than most price signals! That's because the university environment is somewhat insulated from the ordinary pressures of capitalism (and thus better able to exploit young research programmers).
Which means absolutely nothing unless a ton of other people do it as well. A handful of people here and there can be replaced.
More seriously, if you're in basic science, your skills are valuable in transforming the work into a more useful thing to be used later. Using your science factory model, you have created a reusable widget that other people can use. The science factory model does work, you can see its results in things like MIAME: https://www.nature.com/articles/ng1201-365 Where large pooled datasets are used to get insights otherwise impossible.
There's not a ton of low hanging fruit in some fields, as time has gone on the edges are harder and more expensive to see to be at the cutting edge. Ex: you spend $2M on a microscope that does a cool thing and two years later the new model is all that, a bag of chips, and a soda for the low price of $750k. You hope you have a good enough relationship with the vendor that they will either mod or upgrade your system, or that those two years were enough for you to get ahead. It probably wasn't. And you now have a not as fast ferrari for more money than the fast ferrari.
There is a massive glut of international students willing to work for basically nothing, beholden to your PI by their visas. I say this not as xenophobia, but I was the only working class American (my parents do not have degrees) in the department. All students/postdocs that I worked with were from other countries, or if they were American, their families were doctors, or a faculty member. More generally, the kind of people that might own horses :D.
No firm would take this work on, as the profits are not clear, and the time scales for success range from two years to never. In this case success is "great job publishing, we'll give your lab another 2-3y of funding." After which, you better get good at writing books and eating pasta.
Imagine research into a technique for getting better blood pressure readings from people who are so nervous around medical settings that their blood pressure spikes (or more basic research into the mechanisms of blood pressure and anxiety). This is a valuable thing to society (more accurate data informing treatment decisions for individuals, screening for physically demanding jobs, life insurance, forecasting medical spending for Medicare and the like), but it's not worth a lot to anyone in particular.
For the field you described originally, complex imaging devices, there are only so many users of that research so it's conceivable that work could be taken up by a corporate R&D department.
There are all kinds of other very useful research topics that are very valuable to humanity as a whole but it's not clear exactly who should pay for it (I'm not saying you aren't aware of this BTW, hopefully I'm adding support to your argument). In those cases it makes a lot of sense to take a fraction of a cent from everyone and pay for it that way, as we currently do.
We had a couple electron microscopes that we modernized from film, and the companies we contracted with mostly dealt with materials people.
The problem with universities pushing research that clearly has some short-term financial reward (due solely to patents and exclusive licenses under the 1980s Bayh-Dole law) is that they neglect basic research and so close the door to the potential of truly fundamental discoveries like that. This is generally known as the corporatization of the American academic system and it's really been a disaster for basic technological advances.
I am wondering whether we've seen the reverse of the idea I was originally challenging (if research was valuable it would be a business), where universities captured a profitable business because it was being neglected by the business community (and were distracted from basic research).
That theme changed after 2008~ when NIH was flat funded and most universities were gazed upon by the Eye of Sauron for funding. A lot of places that were basic science focused, let's say at the level of studying a set of proteins in mitochondria, had to figure out how to connect the dots to disease or therapeutics. Not everyone made it.
Also, universities got into the game of stacking patents to license. I don't know the arc of that, but I know for sure after 2008 my Office of Technology Transfer was really into it.
Ex before: "We study apoptosis signalling in mitochondria, to understand how mitochondria are related to cell death." After: "We study apoptosis during heart attacks, and how mitochondria contribute to cell death in ischemic conditions."
Something along those lines.
Other good replies here, but this part of the comment reveals some assumptions that need better definition. Having been both, I can comfortably say that academics aren’t “workers” in the same way that industry programmers are “workers”. The parent comment is not correct about the norm; programming for research projects is not usually sold for profit later to industry. It happens occasionally, but most academic work stays academic. Sometimes when it does happen, it’s in the form of a spinoff company that brings the original authors of the software, and so they end up getting some of the proceeds… when the business survives.
Also the top comment didn’t say ‘undercompensated’ - in business this has a clinical meaning that someone is being paid lower than market rates. We know that adademics pays lower, but we do not know that it’s lower than market rates for academics. It’s plenty true in industry alone that you can make a lot of money at Facebook or very little money at a small indie game dev shop. That doesn’t mean the indie game devs are undercompensated, it means they’re in a different market.
Starting firms to compensate researchers better is what pharmaceuticals (for example) are. The problem with your suggestion is that the need for income can undermine the ability to do research that is unbiased, risky, controversial, critical, or just free of agenda. If you pay researchers in line with what Bay Area programmers get, it will put an enormous burden on the PIs to make many multiples more money than their peers, and compete with them using a small fraction of the number of people of peer groups.
The whole point of academia is to subsidize research before it gets to an application phase. How can a private firm compete with academia who benefits from government funding and are tax exempt? Trying to pin this problem on "capitalism" is just lazy.
Neo-liberal capitalism fixes problems?!
I see this scenario described in medical research all the time with people saying that industry just leeches off of academic research and what people leave out conveniently is the vast amounts of money and research that goes into translating research into a real drug (billions spent on clinical trials to meet regulation, millions to billions spent on scaling manufacturing and synthesis of the drug to industrial volumes, drug delivery like pill design or injection methods)
Additionally many industries also do have well-paid research positions that "move the needle" on science and basic research. While they're more targeted at producing and supporting products instead of full liberty to exploring just for the sake of knowledge, it's not like there is a complete black and white poorly compensated academic research vs industry.
Then the competition can come in, i.e. whoever can produce pure preparations of those drugs at the lowest cost will win the most market share. This means investing in top-of-the line manufacturing platforms (much of this is now outsourced to India, Mexico, etc. for drugs being sold in the USA) instead of squatting on the patents, blocking competition, and using monopoly status to jack up prices.
Yes, this would greatly reduce the profit margins and perhaps the stock prices of Big Pharma outfits, but the overall benefits would greatly outweight this. As a practical example, look how the best Covid vaccines (mRNA types) have been monopolized, leading to low rates of vaccination in Africa etc., even though that was technology developed with taxpayer funding at public universities.
It's a misunderstanding of the market to suggest that somehow 'the FDA will lead the trials'. This is about as likely as a manned mission to Venus, it won't happen, and it shouldn't happen for good reason (cost vastly outweighs the benefits).
It's also a misunderstanding to suggest 'whoever can produce pure preparations of those drugs at the lowest cost will win the most market share'. The 'cost of manufacture' is most cases is not a material or relevant issue.
Your example of 'COVID' monopolization is completely upside down - companies didn't maximize their profit potential there, and may not have even developed such vaccines in a normal case, they were giving very special prices to places like 'Africa' - and none of this has anything to do with 'low uptake' in Africa.
Africa has 'low uptake' for the very same set of reasons they don't have electricity, or consistent electricity in many places.
As far as Covid-19 vaccines, there are actually many companies ready to go right to production if those patents are released to the public at this moment, and that would greatly increase supply, and that would benefit the whole world, instead of a handful of pharma CEOs and affiliated shareholders.
That's just conspiracy. Many companies have not been enforcing any patents related to covid vaccines. The reason why the developing world does not have high vaccine adoption is because of their infrastructure, not because of patent blocking.
https://www.theguardian.com/world/2022/may/03/covid-vaccine-...
Of course not all companies are going to release their patents, but the fact there are several that already have, means its not the main problem. mRNA vaccines require extreme temperature controls during transportation and that incurs much more of a prohibitive cost in those countries than any amount of patent royalties do, even if there were no waived patents.
There are cases of huge shipments of vaccine donations going to developing countries that then go unused or underutilized because they do not have the resources to transport and distribute them effectively. Look at the COVAX initiative. The countries that failed to get vaccine rates up even with huge donations of vaccines lacked the infrastructure to distribute them whether it was the temperature controls or not having enough syringes.
"in Benin, only 267 shots were being given each day, a pace so slow that 110,000 of the program’s AstraZeneca doses expired...The vaccine pileup illustrates one of the most serious but largely unrecognized problems facing the immunization program as it tries to recover from months of missteps and disappointments: difficulty getting doses from airport tarmacs into people’s arms."[1]
[1] https://www.nytimes.com/2021/08/02/world/europe/covax-covid-...
- Some type of independent trials might be possible, but there already is a lot of oversight. That's complicated.
- 'Releasing the patents'. I'm sure everyone in the world could release all of their patents for everything, and things would be good for about 2 years, but we'd likely never see another drug produced again, ever.
This completely trivializes and misses the fact that the manufacturing process itself can be patented.
>As a practical example, look how the best Covid vaccines (mRNA types) have been monopolized, leading to low rates of vaccination in Africa etc., even though that was technology developed with taxpayer funding at public universities.
That's just patently false, Moderna for example, has been waived patent infringement related to covid vaccines. The reason the developing world does not have high rates of vaccination is not because of patents but primarily because of their infrastructure.
Almost zero research is commercialized directly, in a manner that equates tech to 'product'.
There are usually enormous costs in applying research to markets - just because something 'makes a million' doesn't mean there were no costs.
As for software:
We probably need cleaner, simpler tools, better SaaS for many things.
We just can't afford to have a lot of devs doing research.
Think about the zillions in lost man hours due to Python weirdness of various kinds. It's a giant productivity sink.
Also, I hope tooling for many researchers starts to improve.
I think the target should be, in most cases, that researchers themselves have the tools available to 'do their work' without having to hire devs.
Multiscale Visualization Using Data Cubes
A while ago I saw a position for image recognition in astronomy advertised by the university of my town. It sounded all very exciting for someone who enjoys figuring out solutions for a complex task -- until I looked up the pay grade: Less than half of my current hourly rate, without the flexibility I have (working from everywhere I want, even pre-Covid). Well, the problem seems to be that a competitive salary for a software engineer would probably have to be higher than for the professor leading the group.
As an aside, I recently had a discussion with a friend in my country's military about the cyber defence forces wanting to recruit software engineers. There is a similar problem here: if they do not use contractors (whom they can pay what they ask for), they have difficulty finding an appropriate pay grade, since a well-qualified software engineer would have to be paid better than a general.
Loved the creative control and influence I could have (even as a mid-20s tech lead on a title), loved my colleagues, loved the work, and even enjoyed the satisfaction from shipping a golden master after a multi-week crunch period.
Ultimately, a hedge fund was willing to pay me a multiple of what I was making in games and I decided I’d rather have a house than work on games.
Look I’m willing to take a massive paycut, I was starting to come around to the time cost, and hell, I’ll even work in Matlab. But:
> you don't have to work weekends, and you also don't have to feed the mice on a Sunday night
This is just wild, and if anything reminds me of low skilled near minimum wages jobs.
From my experience "science" does not appreciate RSEs, the compensation is bad, the freedom is bad, and you get all of the big company bureaucracy from the University.
It's just a bad environment and to top it off, even if you have a PhD many will see you as "less than."
In reality the line is much blurrier. There can be no innovation and iteration without implementation of ideas and the RSE work is just as important.
But unless this view changes, nobody wants to be an RSE.
Is there some way to make human respect feel like it's not a zero sum game? The world may never know.
Respect is non zero sum but status is definitionally a positional good. If you’re number 1 someone else isn’t. There can be uncertainty about status but ambiguity always collapses eventually. Everyone can be treated with respect but there will be a prestige or dominance hierarchy in any group of humans, subtle as it may be.
I am a licensed professional engineer (mechanical) and I work in academia (though my work is varied and does involve some code), my official title (in french) is "Professionnel de recherche", which translates to Research Professional. It is my understanding that in most of the english-speaking world, this position (someone that works for a research lab, who is qualified above the technician level, but not a PI nor a postdoc or student) is called "Research Assistant".
I write "Research Engineer" on my résumé/linkedin, because TBH for most people who don't have much experience in academia, "Research Assistant" sounds like an admin assistant (secretary).
I did tech support at a university for a bit. I certainly wasn't as high status as the professors, but they mostly respected me and the value I provided (especially when you do things like help them recover that next book they were working on, or whatever).
On the other hand, I've never worked in the national lab system. If you're still monitoring this thread, I'd appreciate it if you contacted me at the email in my profile!
I am a mechanical engineer by degree, but a software engineer by 30 years of practice. That has allowed me to thrice straddle the divide where I worked on software that was not just software for softwares sake, but rather as a value add addendum/enablement to an institution whose roots were in more real-worldly devices. I also spent time on the other side of the divide working twice for companies whose primary product was software.
This “less than” is a very real thing. Software is still a very new thing relatively and it’s taken over the world in just 30 years time. The power dynamics, entrenched in generations haven’t had time to rebalance.
I’ve seen this run the other way as well. Where companies that are primarily software, look down on the other disciplines that participate, a necessary evil, but resented.
I have yet to find a company/institution where the trifecta of mechanical, electrical, and software is balanced in mutual respect. If one such exists and you are looking for someone that would love to work in that environment, drop me a line. I’m skeptical that there are any. It seems one discipline or the other always triumphs and tramples the others down. On off axis variant of Pournelle’s iron law of bureaucracy .
I switched from engineering into comp sci at my university and I can safely say that the elitism for software was stoked by the elitism of engineering.
While I was in engineering, everyone bragged about how they were going to get “The iron ring” (a thing we give to only engineers in Canada). When I switched over to comp sci, all my old engineering buddies would say to me, “yeah, but you don’t get the ring” like it meant everything.
After I graduated, I noticed the comp sci “software” folks didn’t like engineering folks much (especially new grads). When I asked why, they shared similar stories to mine. The engineers and their pumped up pride at my local university had hurt their relation to other fields by being arrogant.
I’m not trying to justify this kind of prejudice, but the reasons it happens are fairly obvious to me. It’s sad cause all of us are so similar in our trades and will likely end up doing similar work as well.
The RSE role doesn't seem to fit into their current model. It doesn't work from a career path, nor from a competitive compensation standpoint.
I suspect that the housing crisis is also going to push more people out of these positions.
If you have a PhD in material science, physics, chemistry, biology, etc... and are reasonably knowledgable in Python (or similar), perhaps spend a year as a postdoc. After that year, seriously consider moving to a tech company.
Would love to see more startup medicine development companies ...
Alternatively, would it make sense to provide RSE services as a company, for very high hourly rates, separately from the rest of the university, so status and rules are less of a problem?
1. Benchling 2. Enable Medicine 3. Radix.bio
a. You're implying your colleagues are incompetent.
b. You can't use that money for "science".
c. It puts in painful relief that the software guys are much better paid than you are, implying that their work is more valuable than yours. But a lot of scientists put up with low pay because they believe their work is really valuable.
PhDs having their own version of the code, incompatible with the one of the Postdoc sitting opposite of them. Each slaving away at their niche project but nobody there to bring it all together. This particular university had really great code and they thought they could sell it. But they did not even use git and when I gave a presentation there about it, I was met with absolute rejection. They don't have an idea on how much effort it takes to maintain a commercial code, unit tests, customer support etc.
I'm now working at a company that produces simulation software spun out from university. Fantastic job and we do all of those things mentioned above. But obviously my paper output has been near zero even though we sometimes do cutting edge research.
I wish I had a solution to the problem as so much grant money is wasted that produces a paper or two and the corresponding software for that just goes to digital nirvana, it's a real shame.
What needs to be said though is that reimplementing something from a paper is near impossible. I had to do that a couple of times and I was only fully successful if the paper was accompanied by some open source code as there always are tiny little edge cases or initialization details that won't be covered in the paper.
A bigger challenge is that most PIs are not project managers and have no experience as such, so they don't know how to express what they need in a structured way, or how to steer their group to collaborate properly. Outside computer science, many would struggle to budget software development or compute properly on a grant application (and the assessors have zero idea either).
The scientist just wants to focus on their research and once they have a barely working proof of concept, hand it over to the engineer to figure the rest out. The engineer wants a well specified design and prototype that they can lightly refactor to clean up, scale up and turn into a product/tool.
The reality is that approach makes it way harder for both, though most often harder for the engineer as they are generally at the end of the chain in Academia and have little power. For example, the code or spec from the scientist is often terrible, so the engineer needs to start from scratch and keep going back to the scientist to spec out the design as they were not involved at any stage prior. They may even find edge cases or flaws the scientist had not considered that are fundamentally problematic to turning it into a viable product/tool.
This is why the big corporate/industry research labs often have high level RSE that are involved in the research process and get their names in papers (they sometimes have PhD's themselves). They are not optimising for the scientists time, but for the companies resources
I know some of the big research councils do this in the UK. For example STFC has a program where they'll work with universities and companies to production-ise research code.
> The scientist just wants to focus on their research and once they have a barely working proof of concept, hand it over to the engineer to figure the rest out. The engineer wants a well specified design and prototype that they can lightly refactor to clean up, scale up and turn into a product/tool.
As you say, this is a great idea in principle. In reality I think that it's really difficult to make it work.
[0] https://www.software.ac.uk/programmes-events/carpentries/sof...
That would ensure that the contributions of the PhDs does not get lost and they learn how to properly contribute to a project.
There are of course scenarios where someone comes up with some very high impact work, and there's an obvious need to make it robust or user-friendly, spin-out, etc.
It works less well for groups where everyone works on different or loosely related projects. That's not an efficient use of an RSE's time, in my opinion. Though of course you can have a situation where lots of people do random projects using the lab's core code. In both cases, there is a use-case where RSEs embedded in a university can train students on good coding practices.
> RSEs should not be writing code for students doing PhD level projects in my opinion
So should a mechanical engineer PhD be designing and making all their own robot parts? Or should the shop engineer help them? The few mechanical engineer PhD's in robotics I know made a few early prototype test parts themselves with help from the shop engineer, but the shop engineer made and even helped design most of it, especially the final prototype.
> As you say, this is a great idea in principle. In reality I think that it's really difficult to make it work.
The point I'm making is that it does work and its proven to work very well (which is why the major industry labs do it). In my experience its Academia that doesn't like it. Anything which appears to take power/freedom away from scientists and gets in the road of their research is rejected. Though I think the core reason is (as other comments have mentioned), there is no incentive for Academia to make it work. The funny thing is that having a RSE working with them would actually help the scientists in the long run and allow them to focus more on the research because they wouldn't have to do everything themselves.
I would argue these should be included in best practices for software engineering.
> So should a mechanical engineer PhD be designing and making all their own robot parts? Or should the shop engineer help them? The few mechanical engineer PhD's in robotics I know made a few early prototype test parts themselves with help from the shop engineer, but the shop engineer made and even helped design most of it, especially the final prototype.
This is an interesting example. Every mechanical engineer I know has huge respect for their in-house machine shops. Everyone has a story about some design they submitted for fabrication, only to be told by the machinist that their design was terrible and they should do it another way. Generally machining jobs are very well-defined though, you have to submit CAD documents, tolerances etc.
The shops in universities I've worked in have a strong incentive to help people optimise designs because they're the ones doing the manufacturing, and they know what sort of things will work and what won't. But by and large this is informal. Usually this comes in the form of "have you thought about designing this another way, because this is really difficult/expensive/time-consuming to machine". Maybe this is just a cultural thing for machinists?
The PhD question - if your project is to design a new type of part then you should probably do the design. Should you make it though? It depends if the project is specifically looking at fabrication. Otherwise it's normal to dispatch this to a workshop.
In my opinion, it comes down to what your PhD is training you for or what you're hired to do as a postdoc. If your job is data analysis, then I think you should be writing code, but you should be able to get guidance and support. If you're a field biologist with no coding experience and you want to develop an app to take measurements, then that's a case when contracting it out to an in-house development team makes sense. I'm not saying it can't work, but the make in making it work is important.
If you incentivize RSE's properly then their time will become expensive and we need ways of figuring out how to maximise their impact.
And what if the PhD’s output relies on software to even exist?
I'm saying this as someone who did a PhD and who wrote a lot of code, including refactoring legacy codebases in my group.
Where the utility in having an RSE lies is where the group is all using some shared codebase that gets touched by everyone. This is the sort of cruft that I had to work with: legacy frankencode that generations of students and postdocs had added to. It would have made a ton of sense to pay someone to spend a year optimising it (which is ultimately what I did). But you want to make sure that RSE's maximise utility in the group. Having them work on individual student projects is not an effective use of their time IMO.
Also from experience, it helps if universities have closer ties to industry. As an industrial player you can't work with software that does not deliver reproducible results.
> containerisation
Pick one.
Also, I'm currently a postdoc in medical imaging at UCL, super interested in learning a little more about the group in Oxford you mentioned if you're OK with sharing a link/group name? I may be able to guess but just want to check!
They are almost always time and resource strapped, so it becomes a near necessity to deprioritize factors like readability and maintainability.
Still, many RSEs could immensely benefit from applying basic best practices.
From the perspective of the university and the PI, they can just get a graduate student to do the work for 1/5 the cost.
That is what they do. However, some projects grow in scale where a grad student cannot handle it well, or if they try, it will be a detriment to their research. To give you an example, only one of the PhD students (non-CS engineering) in my group and the ones around me had taken data structures or algorithms - yet everyone's thesis was "computational" (i.e. numerical computation, etc). None of the advisors would appreciate their students going off and taking serious CS courses.
When a research team wants to go to the next level, they need to hire someone with better SW skills.
Why would anyone want to do that, when a six-figure tech company salary is just next door?
First, if you choose to work for a non-profit, you should expect a huge pay cut. That's the way the society works. Or actually it's the opposite. If you work in a for-profit role, the society allows your employer to keep most of the value you create. If you create a lot of value, that gives you the leverage to negotiate much higher compensation than your peers in less profitable roles.
Second, like in any business, you should focus on what brings the money in if you want to advance your career. You should be the PI instead of the lab technician. Instead of building the software other people ask, you should become an expert in the field and build the software other people will need in the future. For example, bioinformatics as a whole used to be more of a support role, but today there are many high-profile bioinformatics PIs.
Third, many people in the industry fail to realize that you don't really work for your employer in the academia. You work to build a reputation for yourself, which makes you valuable to the employer. If you want to continue working for other people, the upward career path from support roles goes to the administration.
That feels like how it works outside of academia as well.
In the industry, your employer tends to own the projects you work on.
It may be the culture in some ecosystems but it's entirely false. A nonprofit can pay competitively with organizations of its size. I've worked at nonprofits and I've seen this line towed far too many times as an excuse to lower labor costs for people producing value for the organization. If your nonprofit is service oriented in anyway and requires people, appropriately investing in those people through their compensation and WLB is an appropriate investment.
In theory, a nonprofit should be able to outperform and outpay a for-profit entity of the same size, largely because there shouldn't be an expectation of the organization to cut a big slice out to investors but instead to invest back into itself. This could mean increased hiring, grabbing top talent, investing back in a cause and so on.
Very often, the level of funding is what it is and the non-profit has very little control over it. If you choose to pay higher salaries, you get less work done. This is common in the academia, which is usually funded by taxes and tuition fees.
Sometimes there is even an inverse correlation between funding and salaries. If a charity chooses to pay higher salaries, its "administrative costs" increase. Donations may then dry up, because people consider the charity inefficient.
Many non-profits rely extensively on volunteer labor. Salaried professionals often contribute their time and expertise to a worthy cause for free. That puts paid employees in an awkward position. It's hard to argue that you deserve a higher salary when the market rate for your services is 0.
Inside academia and outside of it.
Inside academia you are not on a tenure track or similar, and will have to put up with a lot.
Outside academia, your peers will be making 3x or more than you, working less hours, with less stress, etc.
The reason RSE's jobs are hard to fill and often aren't even opened is that they don't make sense. If you are good enough for an RSE job, you will be good enough for research postions at FAANG. Those pay 10x more, so you also need someone willing to not accept that 10x pay, and also willing to work double the hours.
RSEs making a reasonable pay for the skills they require make no sense either, because that would put your pay at 2x that of professors, etc.
On that note I wish levels.fyi had RS and RSE roles...
You don't have to be at a "tier 1 paying company" to exceed 320k TC, that's easily achievable at any tech company that has gone public in the NYC metro or Bay Area. TC at a FAANG (MANGA or whatever they're calling it) for a more senior IC role can easily cross the $500k mark. If you've been their awhile and have been accruing shares that have increased dramatically in value, crossing the 7 figure mark is not unheard of.
So you don't have to stretch the facts too much to realistically achieve the 10x. If you take two very qualified engineer graduating with a PhD, one chooses to go to Google and stay there the other goes to a university and works as an RSE. Fast forward a decade, you'll definitely be seeing a 10x difference in total comp.
Poor compensation: Yup, I'll cop to that. I could be making way more in "industry" but I make enough to live and I don't build ad-ware and do take pride in my work.
Being a PhD but little respect: I actually have the opposite experience. I'm not a PhD, I don't even have a Masters. But I can write decent software and I get respect and recognition from my PIs and other top-tier faculty that we've collaborated with. I feel like I've gotten just as much respect with my humble BS in Comp Sci from a state school than if I got a MS/PhD.
Micromanagement: I have the total opposite experience. I've worked with the same group for a long time, maybe that changes things. We apply for and receive grants to work on projects. We make high-level assertions for what we will do during the course of the grant and our PI gives us lots of lee-way to meet the goals of the grant in whatever technical way gets results, is FAIR, and open. Granted, I may have just lucked out here.
(My research institute didn’t have a concept of “research software engineer” so I was paid as a generic research assistant.)
B.S. in Physics (w/ years of significant undergrad research), but worked with PhDs+ in national lab scenario. At some point everybody assumes you have an PhD.
Little to no micromanagement. It was more professional scenario.
IMO an RSE at a mid-ranked state school hired by a PI and embedded into a research group, compared to an RSE at a top-tier research University - at a lab or institute that employs many engineers, they will have vastly different experiences and probably pay as well. There's exceptions to the rule, but this is generally my experience.
There's a good chunk of both jobs, but I would not work the former.
My SO is a research software engineer.
1. As you say, she doesn't build ad-ware and do take pride... 2. relaxed and convenient job timings 3. In EU - close to 31 days paid + usually Professors are very generous about going home early etc... 4. Pension, unions, unlimited contracts 5. No one gives a shit when github/gitlab/heroku loses credentials on a Easter weekend. (We were in roadtrip) - chill life.
Instead I got given an intern from the internship program. While I tried my best to give them a good experience, let's just say that wasn't the boon to productivity my manager thought it was going to be.
To make matters worse, one of the researchers in the PhD program had been told these changes would be ready shortly before I even started, and the PhD candidate had made a series of choices about their studies relying on those changes being available. Felt bad to have to break the bad news to him after the lab manager had been blowing smoke up his ass for so long.
When you put it that way, it makes it so clear why nobody wants to sign up
(BTW I'm the same way, would love to work on research, especially if it is something which could improve people's lives, but I'd have a hard time switching for these reasons)
Definitely true.
>have zero autonomy
Might be true? But in the right research group, this is definitely not true. I actually find that I had far more autonomy in academia than in industry. There are far fewer deadlines, less time pressure, and more curiosity-driven projects that are entirely owned by you.
>and get no respect
Might be true, but I find a lot of people in industry feel they get no respect as well. As an academic, I actually found that I respected academics more than I did people in industry. Though this is entirely anecdotal.
I was an academic until my 30s before going to industry, and actually find that the autonomy is the number one reason people like academia. YOU drive the projects because only YOU know the research that deeply. YOU get to be at the forefront, and YOU know things nobody else might ever know, until you get to share them through publications and presentations. Academic projects have essentially no deadline (projects are on multi-year long timescales rather than quarters or months in industry).
Yes definitely true, a junior software engineer in industry earns about as much as a Professor in academia. How would you justify such a salary at the university?
> have zero autonomy,
I've written that before. I would argue it's the complete opposite, autonomy is the one big advantage of academia. If you are a software engineer you likely have even more autonomy than most others, because your work is not directly linked to conference/grant deadlines etc. typically you would be given very rough guidelines on what you should try to achieve and then left to your own devices for long periods of time. Most PIs would neither have the time or expertise to micromanage a software engineer.
> and get no respect.
Again not my experience at all. I know that in my area people would respect a software engineer. I know when my partner was working as a research assistant in a medical lab, I wrote a small script to process some huge spreadsheet that they always processed by hand (work that took sometimes several weeks). They treated me like a god and even wanted to include me on the paper. The thing to remember though is you will not be considered a researcher, so likely not get grant funding and your job progression would always be in more technical roles, not in the academic track. I still don't see any difference in respect.
Respect could mean different things to different people, but I find the things you wrote not to be very respectful, but being nice in my evaluation.
I have worked in a research lab. While the folks I worked with respected my work, what I would actually consider respectful is including me in the design of the work, not giving me my part of work after they designed the thing they want to work on. And while most job industry is not good in this respect, working in research lab is worse than that. Almost all research labs treats SWE like cost center[0].
[0]: https://www.kalzumeus.com/2011/10/28/dont-call-yourself-a-pr...
> I have worked in a research lab. While the folks I worked with respected my work, what I would actually consider respectful is including me in the design of the work, not giving me my part of work after they designed the thing they want to work on. And while most job industry is not good in this respect, working in research lab is worse than that. Almost all research labs treats SWE like cost center[0].
I don't quite understand your point, do you mean that you as an SRE want to be involved in the grant being written? That might be appropriate if the grant is directly on the topic of software, but completely inappropriate if the SRE writes the labautomation software to be used in the grant. So if the researcher asks you to write a GUI to some lab-instrument for example is that being disrespectful?
Regarding your citation, I also don't get it. The text says that MBAs consider SWEs as cost centers, so that implies it's worse in industry (there's very few MBAs running research labs). I can also tell you that academics are largely considered cost centres by university admin.
The right way to think about this is in terms of long-term politics and the success of nations. If you disincentivize national research, your country will fare worse than its competitors. For me, that's the US- if the US doesn't turn around the incentive systems for doing high quality academic research in the national interest, we will eventually be at an economic disadvantage to China.
I've literally sat in meetings trying to get access to scientific projects I'm interested in and applied for jobs even at the university I already worked at in similar capacities, and yet, it's filled with PIs just wanting grunt labor (i.e., what they're used to from graduate students) to build things that they understand or you absolutely never hear back or the pay is a third of what you already make. And then there's pockets of software people in this niche world that only want to do things their way, because that's all they've ever done, and they can really be quite obstinate. They're like this massive pillar you must somehow work around.
I really, really, really like writing software in scientific contexts, but it's such a quagmire. I'd like to return to that world some day with the lessons I've learned outside of it. Software in support of an interesting domain or context is just really fun. I've also tried getting a job at an architecture firm that had some very neat ideas for building out a tool, but the response was basically the same as described above.
These people desperately need software help and have worthwhile projects that need software help. They also need to learn to collaborate over just parsing out obscure goals and being upset when what they get matches their three sentence email and not their vision in their head.
If the requirement is that it lasts 30 years, you can choose more timeless languages like C, over say Rust.
And yes, way too much research code reads like somebody's first coding project.
Moreover, people tend to forget that scientific software evolves much more slowly than commercial software and the number of people who work on this are much fewer, so why always adopt the newest method, if the code is still working. Show me commercial projects that have had the stability of something like LAPACK or BLAS, those essentially had a stable API for > 50 years now, with very few bugs.
But to be honest, I've seen this in industry as well from software developers. So maybe it's just a law of software development that's amplified in scientific software.
I can definitely tell you that if I talk about my work non-scientists become quickly uninterested.
This is all painting with broad strokes, but in my experience, the most interesting intellectual conversations I've had have been with people with the least amount of degrees. I always feel when I talk to someone with a Ph.D., they're afraid to let it be known someone might know more than them and if that happens, then it's suddenly something that's not important.
The book Disciplined Minds gets at this some. It's part institutionalization and other things. My gut feeling is that it's exacerbated by publish or perish culture and the push to always be right. So people that are brought up through that are intellectually risk averse.
> In my experience, scientists do not like anything that they perceive as "soft". I like to say that the hard sciences are actually the easy ones and the soft sciences are actually the hard ones. (As it turns out, Herbert Simon also said this.) The point is that when you discuss things like design, complexity, architecture, systems thinking, etc., I've found that this does not stick, so to speak. Scientists do not see beauty in form equaling function. In my experience, scientists like learning about things they consider science-y enough for their tastes.
In my experience what you say is true for engineers (and software engineers are amongst the worse in this respect), but much less so for scientists from the hard sciences (physicists, chemists, biologists). You can also see that from many famous physicists being involved in philosophy, while many engineers consider philosophy "irrelevant talking about nothing".
> This is all painting with broad strokes, but in my experience, the most interesting intellectual conversations I've had have been with people with the least amount of degrees. I always feel when I talk to someone with a Ph.D., they're afraid to let it be known someone might know more than them and if that happens, then it's suddenly something that's not important.
Maybe we converse in very different cycles, but generally for people I know the "higher up the foodchain" the more they ask even very basic questions. In my experience you can often distinguish who is the most accomplished scientists in the room by listening who is asking the most basic questions.
All this is obviously anecdotal evidence, but is reflected also by famous accounts like e.g. Feynman's account of the challenger commission.
As for "breaking in," I dunno. The hiring process seems very fair and international (very few locals work there in a technical capacity AFAICT), and there are plenty of people with flexible ideas. Not everyone, for sure, but then there are so many different groups there you couldn't possibly blanket everyone with that statement.
Tldr: They view software as just a tool and don't see how they'll benefit from following best practices.
This is not really the core issue here, but yes, I've encountered this many times, and I have to admit that in some professional contexts I use "codes" as a shibboleth to identify people who aren't really in touch with the software engineering world.
The reason why they can't and don't want to hire software engineers to build well designed software is because they don't have any money for it and it is very mismatched with their incentives. If a PI gets a grant for a project, they typically get a constrained budget to achieve the scientific goals of the project. There is usually no money (and it would not be approved) to engineer well designed software, because that doesn't produce scientific outcomes.
Now in the situations when there is budget for this, which are the super large institutional research initiatives, like e.g. CERN, they build incredibly well designed software. The SKA for example will have the largest storage cloud in the world.
If one wants better designed scientific software one needs to seriously overhaul the scientific funding process and quite significantly increase funding.
They/We have probably been calling it that since before you were alive. I wouldn’t be surprised if it comes from the 60s or 70s.
Every group of people has their own terminology for stuff. There is no right or wrong.
Why not? "A code" is a software package which solves some kind of problem computationally. Plural, codes. I'm sure it's been in use almost as long as the programmable computers. Here's[1] a link to an article from 1968 which uses the term in the singular ("The MENE Neutron Transport Code".)
[1] https://www.osti.gov/biblio/4819611-mene-neutron-transport-c...
Then you probably haven't been around physicists, or read enough physics-related material. A non-scholarly reference which can't be easily checked online is George Dyson's "Project Orion," where Freeman Dyson (author's father) is quoted as mentioning this or that "code" as significant for the design of various Orion parts. (That should be a native enough English speaker.) The time-frame is late 1950s to early 1960s.
That reference you give doesn’t appear to be available online but I’ll take your word for it.
I think the point is that the physicists were among the very early adopters of computing systems. I’m not sure about the history of the term “code” used this way, but it might literally predate compiled software since Ulam and von Neumann invented the Monte Carlo method in 1948 during the Manhattan Project.
Personally I think it just sounds outdated now, but it’s pretty common to hear people call their software “a hydrodynamics code” or “a molecular dynamics code”
The tendency is to call computational science software “codes”, but I’ve never heard a scientist call, ie a web browser, a code.
There’s a different connotation. Like how theres a different connotation between a program, an application, and and app. The are largely synonymous, but have slightly different uses.
Obviously, what word they use for it is arbitrary but it's indicative of an actual problem - they're all totally cut off from the software engineering / CS community. Lots of them think programming is like learning to juggle or ride a bike or something, a skill that anyone just learns and then they're done. Then, obviously, they make major mistakes that invalidate their conclusions and nobody notices or cares because it's all student code, all the time.
See for example Code_Aster (as in, "a code"), which is one of the largest projects for open source FEA (and it mostly rooted in academia).
As someone with a mathematics background, whenever I encounter someone in another community using mathematics (like economics or accounting or finance or whatever) using mathematical terminology that seems silly to me, it's a bad sign.
...please excuse this shameless plug: I'm interested in people who would like to work with me to build research software (both on-premise and AWS cloud) using best practices, in particular using Rust. If you are interesteed in machine learning, natural language processing and information retrieval and would like to work with me as a research software engineer, please do get in touch (leidner at acm dot org). Our new 2 TB RAM dev research group server running Ubuntu 22.04 LTS is waiting for you (CPU/GPU cluster ressources also available) to set up (a VM for) your CI infrastructure. Your software will be used to push forward the frontier of research in machine learning and its applications to language (summarization, question answering) and information retrieval (vertical search engines, learning to rank, unsupervised topic modeling).
Edit: academia can be poorly funded and bureaucratic. The former can be fixed by grant applications, the latter must be tolerated until politicians/activists fix it.
The other comments in this thread indeed point out the issue of poor salary compared with industry, this is a problem that I feel will persist. Historically, the benefit of working at a University over a company has been greater freedom, better benefits and (at least for me) the knowledge that you;re working for the public benefit. Unfortunately, the benefits are being reduced as time goes on as Universities are under greater financial stress.
My role as an RSE at the University of Bristol has been in a primarily teaching position for the last few years, where I have been training PhD students, postdocs and research staff (all the way up to and including professors) in both the basics of programming but also testing, version control, profiling etc. My role is funded indirectly by the government as it is recognised by the funding agencies that RSE skills are essential for modern research and they are putting their money where their mouth is by funding more software projects, fellowships and training programmes in this area.
If you think that you'd like to know more about getting involved as an as RSE, or if you work as one an would like to meet like-minded people then you can join as a member of the Society, join our Slack channel or our mailing list (https://society-rse.org/join-us/) or find out how you can get involved (https://society-rse.org/community/get-involved/).
Money. It’s all very well being an RSE but there are two ways of being funded (at least here in the UK). One is centrally funded by the University - universally this means being shoved into the IT hierarchy which is not a good fit. The other is on an ongoing basis by doing 1-2 year contracts. So no job security.
The salaries for both of these are very poor. On leaving I took a 35% pay rise and do basically the same work but in industry. The only difference is that instead of working for engineers in academia, I do work for engineers in companies. Since I jumped, I negotiated another 20% pay rise. If I’d stayed, I’d have got a 3% pay rise. That’s not including the fact I get bonuses, private health insurance, don’t have to pay to park at my own office…
The issue is that the skill set demanded to be an RSE is highly desirable in industry - able to write code well, often mathematical, familiarity with much of devops type work.
On top of that, researchers don’t truly value your contributions. They don’t want to put you on papers that couldn’t have happened without your work. On top of that, they’re hugely demanding about timescales that are unrealistic. I got put onto one project which had contractual commitments to deliver to a major governmental agency to run a live service, and the guy running the project handed me a 5000 line Python 2.7 script that wouldn’t even run and basically said “get it working”. He couldn’t even supply the input files needed to run it for 2 months after that, just ignored emails. I quit before it got any further.
Edit: forgot to add, my boss was a grade higher than me on the national academic salaries scale. He had technical responsibility and also had to line manage 10 people. Even if I could have gotten promoted (impossible - no budget), I didn’t want the line management responsibility. I earn more than the top of that salary band now, and line manage nobody.
In my workplace, I am the rare one who can code C++ to work with drivers for real time stuff and develop C# GUIs for convenience and automation. Everyone else is happy with Python or Matlab scripts that are only usable to one person and always need to be edited even for the simplest parameter changes. I would do analysis stuff in real time in a camera feed and adjust the experiment in real time, while others could only analyze an image after being saved. The difference in productivity is literally a hundred times. In most cases the job would have been impossible otherwise or at least painfully inefficient.
Unfortunately, research leaders cannot tell the difference (or pretend to), and just label everyone the same. They are not software engineers themselves, and are probably even insecure. I feel my opportunities and potential are not being recognized either. I also feel that they don't respect that software development needs more time as they will just dump as much tasks to me as the other non-software engineers.
I have to eventually leave such groups as I don't see their willingness to step up in their coding skills despite being heavily reliant on my output. Draconian intellectual ownership rules don't help either.
My pay was almost doubled (after counting bonus) and with RSUs it’s double+.
I could have squeezed out another $XXk maybe when I left the national lab system, should I had stayed.
The biggest reasons I left were as follows:
1. Too much legacy responsibility. Legacy responsibility is usually underfunded or unfunded but “important”. It’s a lot of work to keep the lights on. NASA usually handles this with contractors which are usually paid even less than a National Lab person.
2. Bifurcation at the lab towards very large projects or very small projects. Not a lot of stuff in the $250MM to $750MM which is large enough to have an effective team and go full time, but small enough to not be anonymous/matrixed out to subprojects. The high end is a very funny place
3. Related to (2) is not enough new blood or turnover. This may be a good thing but you can be hired as a junior engineer and end up as a senior engineer without anybody below you for years. Getting an intern is a dark art if your not a staff scientist
4. Institutional BS. Looking at you, security drones with professional certificates running qualys/nessus scans. and you too, data center people afraid of the cloud.
Some to zero agility in terms of compute - everything is a nail and slurm is your hammer
The university-based RSEs complain a lot about not being respected/publishing. I never really had that issue much specifically, but that may be because there’s a stronger understanding of value in the national lab system. On the other hand, poor leadership/project management means you may still be at the whim of an idiot, servicing their desire for some crappy web app project. In those cases, it’s best to ignore.
I will probably go back after some time when I’m throughly annoyed by BigCorp. In the mean time, it’s fine on the other side and at least I don’t have to patch Jenkins every two weeks.
My PI flat out said when we were trying to hire a python developer that they simply could not be paid 2x more than anyone else because it would cause major issues with team cohesion. 2x is the bare minimum to even get close to market rate for a full time developer who won't dance out the door a year later.
By this logic either everyone in science needs to get a massive raise (stop to the PhD slave labor?), or rse's really have to justify their salary, and when everyone else works 60-80 hours a week for half the pay and is also defacto on call, that justification looks like zero work life balance.
Talk to a funder near you, let them know that congress needs to provide more money for basic research. Be sure to sell it as training future employees for big tech too since they will all eventually leave after you finish teaching them.
They do not have to pay the software engineer anything if they can find one who is happy with getting nothing.
Salary is always going to be a factor in who they can attract, especially when their competitors can offer much better compensation. If they are satisfied with the software engineers they are currently getting, then they can keep doing what they are doing now.
What the professor gets paid matters very little compared to what the candidate can get at a competitor.
I was not complaining about developers being too expensive, I was responding to people who said that they would like to work as a RSE but salaries are too low and explaining why the salaries are so low.
This does play well into my argument that if software salaries in general came down then the rest of the world could utilize software development more broadly across industries, rather than concentrating so much talent into one inflated industry niche (the tech industry is that niche).
That won't work. Academia isn't under-funded, especially not in the USA. The NSF budget last year was $7.7 billion so money is there.
The reason salaries are so low is because academia will generally choose to expand the empire by creating new projects and hiring more postdocs over raising salaries of their existing team. And they can, because lots of people have stars in their eyes about science. They see it as more virtuous than mere product building and are willing to put up with a lot for the associated prestige. Also they may have picked their vocation early on in life when their understandings of real market rates was poor, then get stuck in it.
Some other issues include:
1. Knowing that software development can be a black hole. Nearly half a century after The Mythical Man Month, management of software projects remains an unsolved problem. And we're not gifted managers to begin with. When I do it myself, if nothing else, I have a pretty good sense of when it will be done.
2. What to do if there's not enough programming work for a full time programmer, and that person doesn't want to work on other things.
3. It seems like programmers with domain skills related to R&D, such as math, are in particular demand.
4. It takes a certain temperament to hang in an R&D setting. A lot of engineers just hate it: The rapidly shifting requirements, and the knowledge that something they make will only be used once, or even wrecked.
The dirty secret is that the same programming work that's a step down for a commercial developer, might be a step up for a grad student or researcher who wants to learn a marketable skill. Also, a multi-disciplinary team, including people who can program as needed but also do other things, can be quite agile.
My team has a software engineer, but he doesn't do very much coding for us. He has actual R&D projects of his own. Yet he can help us at a higher level, for instance giving guidance on how we can write better code, and make better use of tools. This brings us a lot of benefit without wasting his brain cells on mundane coding tasks.
Initially my supervisors were very skeptical and thought I'd waste my time, as they did not see any value in software development. Having a programmable and well organized software framework was absolutely instrumental for my work though, and I think I couldn't have succeeded at my experiments (demonstrating quantum speed-up for a simple quantum algorithm) without it. Today the software is still in use in several labs and as far as I know they also hired a research software engineer to keep developing it.
Still, the situation you describe is unstable because as soon as the budget starts to look uncertain that engineer is likely the first to go. I really wish institutions would take this problem seriously and set up e.g. a department wide development group, which could help build and maintain stuff like that and ensure continuity. Unfortunately this seems to be against the current trend of cutting costs. For example, at the university I used to be, they cut the small common electronics shop, which was a great resource since the guys there knew all about who had what equipment, kept spare parts from broken stuff, etc. But in the budget it looked like pure cost I suppose. It's hard to measure the value of common resources and nobody wants to pay.
Software engineering is a continuous effort. Maintenance often requires more time and resources than the original development. It's more like curating a library or a database. Their product is perceived to stay the same. Incremental improvements do not advance the career of the maintainer at all. Researchers who sink their time into software development mostly hurt their academic careers. Even if they publish a paper about ResearchTool 1.0.0, they won't be able to publish another one about ResearchTool 1.0.1 and another one about ResearchTool 1.0.2. Meanwhile, their peers will probably publish "A", "B", and "C".
Science needs a career path which does not depend on the number of publications.
Over the last 2 years I've interviewed at multiple universities for Level A (Associate Lecturer) and they all had similar requirements that I've summarized below:
> you must have completed a PHD. > You must propose and perform original research, assist in upper level academics research, and publish X times per year. > you must participate in fundraising, complete your own grant applications and assist upper level academics in their fundraising & applications. > you must supervise and advise X number of postgraduate students thesis/dissertations. > you must teach both undergraduate and graduate level classes across 3 trimesters. > the position is fixed term for 12 months, rather than ongoing/permanent, with no job security and you are required to reapply each year.
Most current academics I know from Level A to Level E, say they don't particularly want to teach, and just wish they could focus on their research and fundraising to support their research. I met other applicants at networking and alumni events like myself who would like a teaching focus, rather than a research focus. There's a supply of "specialists", but an insistence on "generalists". As the fixed term aspect tends to apply to research fellows as well, the risk from the lack of security combined with the extremely low pay makes it very difficult to enter, and remain, in academia. I can afford to live on $65k a year, but the possibility that every year I may be faced with periods of unemployment and the requirement to relocate thousands of km away makes it a challenging prospect.
This shows the attitude that some folks in these places have - what you do is easy, and rather than your industry being exceptional as a profession that makes its tools open and shares knowledge openly, you are at fault for releasing things which I can get wrong.
"Scientists" in this view are smart people, and anything you lesser mortals do should be obvious to them immediately, or it must be your fault. Training and experience be damned.
I'd hate to work anywhere where that attitude was prevalent, or even really existed at all. And yes I realise that non-software folks often hate to work with us because of the exact same thing!
http://blog.khinsen.net/posts/2020/05/18/an-open-letter-to-s...
There are links back to other parts of the story in the text there. Effectively, they built a model that predicted some real worst-case stuff, and used that as the basis for advice they sent out to try and affect government policy in a bunch of places. When the code was finally released and examined, it was in a bad state - 15k lines in one source file, race conditions, single name globals reused all over the place. The results were trash. Then the blame started to fly.
In this case, due apparently to arrogance, neither was produced.
PhD students are the backbone of any research project. They interact with postdocs and with undergrads in the research ecossystem. If anything, they will be required more in the years to come, not less.
When there's no technical debt, sure, grad students are the backbone. They can happily go about their projects and supervise undergrads, while the postdocs supervise them and apply for jobs.
In a place where there's a lot of technical debt, a postdoc is going to be far more productive than a grad student. In some experiments many grad students spend years floundering around before they learn enough to cut through the cruft and do real research. It's pretty silly to fire all the postdocs just when they've managed to understand everything and replace them with new grad students.
The tricky thing is keeping enough of the technical staff around, and coming up with a system where they are rewarded for eliminating debt, rather than penalized for making their own hard-earned skills obsolete.
I think, like many things, the buck stops with academia itself, its metrics, demands, and incentives. We need more research engineering/science about science academics within compute-intensive science departments themselves. Things like JoSS [1] or Scientific Data [2] are awesome first steps at addressing this.
I have an industry job with good pay with a side of academic work stemming from my PhD. My industry income is literally more than 2x the salary of my academic PI, who holds a senior academic post. My pay, which is probably around the median of most HN readers, is slightly more than the top peg of the academic pay scale in my area - i.e. University senior leadership team. So I do the academic work because I love doing the academic work. I cannot afford to take a full-time academic job, although it's been offered.
I am baffled as to why academic institutions aren't failing to recruit. If they were businesses offering half the going rate, they'd get third-rate engineers, if they had any applications at all. Instead, there seems to be a long queue for academic tenure - and even for non-tenured positions. The only competitive advantage I can see is the sense of academic 'freedom' that these positions confer.
But - it's not freedom. My academic colleagues spend hours each week chasing funding, filling in grant requests, attending meetings. Their teaching load is allocated to them without choice. They're in the office 9-5 same as everyone else. Academic freedom is anything but.
I am lucky that for the research I do, mostly behind a laptop, my industry pay lets me afford the things I need to do it - cloud services, international travel, software licenses, article processing fees. I can 'do science' the way I want to without begging anyone for cash or going to grants committees. And I'm hoping this brand of 'citizen science' becomes the norm, because University-led research is going to get harder and harder the more the gap between industry and academic pay widens.
Because academics (hi!) aren't in it for the money. If you value making lots of money, or having work-life balance, then academia isn't for you (nor is social work, teaching, nursing, fine art, etc.). But this probably makes it hard to hire RSEs who aren't academics masquerading as software engineers...
> I want to see RSEs as equals in the academic environment. Software runs through the entire research process, but professors tend to get most of the recognition and prestige. Pieces of software can have just as much impact as certain research papers, some of them much more so. If RSEs can get the recognition and rewards that they deserve, then the career path will be that much more visible and attractive.
As a former RSE I think this is unrealistic. Academia just isn't structured to value non-professors (it's far from just RSEs that are undervalued.) For me it wasn't so much the money but the refusal to sponsor permanent residence. When you see your amazing colleagues leave due to dumb departmental policies you realize you need to plan your own exit. I miss working in a team with a mix of scientists and engineers but I don't miss working for a university.
No, scientists need to stop hand-waving away everything that's not their area as "trivial and not worth talking about" because yeah, I've heard that exact line from three separate individuals. When I shrugged it off and still produced tools for them that shaved off a literal 1-2h out of their day and giving them more time to do the actually creative things... they were so overjoyed they forgot to say thanks and immediately proceeded to act as if the tools were always there. They were extremely careful not to show any attribution or even basic human gratitude. Would it kill them?
Yeah, frak that crap.
A lot of scientists become severely tunnel-visioned, arrogant, conservative, downright disrespectful, and very hard to work with. Pile the extremely subpar payment and you really have to just be a bored college kid (or an adult with a huge financial safety net with a passion for the area) to engage with those people in a professional setting at all.
I won't go into details about why I think many are like that because frankly, even if I was 100% correct (and my analysis is not at all flattering) it still doesn't matter one bit. Lasting change comes only from within. Scientists should start holding each other to a higher behavioral and collegiate standards and expel those who don't comply. Peer pressure works but many are so independent and without any oversight that they gradually fall under the illusion that their insufferable quirks are a personality or a personal touch that helps them excel in their area (and it's often the case that they don't excel in it).
Looking down on people because they don't know your extremely niche physics research -- that has zero experiments attached to it, by the way -- is not helping your cause with people who usually get a lot of money AND respect AND a better working environment for actually improving your workday. In most cases these people will of course go where they are at least not looked down upon.
Tbf I can replace scientist with developer and it’s hold just as true.
I'm 42 and I've had colleagues at 28 year old that already act like they belong to a council of elders. Made me super sad... they are way too young for that!
"Most RSEs have a PhD"... talk about selecting aggressively for people with no experience building functional software. Hiring only PhDs to write research software smacks of hiring only white males who can drink and play golf to function as business executives. You're selecting for ability to posture, not deliver value.
More important than having one's own PhD laser-focus would be an ability to communicate with a variety of people doing their own research, and understand their domains just well enough to write/fix/improve their damn software and maybe even publish it (semi-productised) as open source too.
A couple of application languages, some shell skills, maybe some data wrangling, and some software development common sense borne of experience.
Having a PhD (in a relevant field) gives you domain knowledge as well as knowing where the pitfalls are, and a PhD in general will mean you're at least familiar with what questions to ask/who to ask those questions.
Where do we sign up?
If anyone has ideas, I'm all ears!
I feel like my experience would be very different if I had one full-time research engineer who was paid a respectable sum to just... help me write code. I like working with undergraduate students, but I can’t do expect them to commit to the task like I would, or a full-time employee. They just don’t have the time.
I’m doing my work extending / improving open-source CPU simulators and it’s just such a huge amount of code. I feel like my work would be so much more useful to others if I just had another pair of hands to work on docs, do ops work, go back and forth on decisions. My advisor doesn’t have the time for such things.
- - -
On the topic of academic pay: as a PhD student in Boston, I get paid $2000 a month pre-tax. Studios and 1BR apartments renting for under $1500/mo in a 30-minute commute radius of my campus are basically non-existent, so roommates and/or long commutes are a must. I’m realizing now why so few US citizens participate in graduate CS/CE programs: if you aren’t looking for a path to citizenship, it just doesn’t seem like a great way to spend your mid-to-late 20s.
RSEs are a different breed from SEs from industry, the work is not about scaling to a billion people but being nimble and adaptive to new results and ideas.
If one wants to see change, then hire senior PI positions who do research software engineering. This will help create a hub and prevent the feeling of being a “pet” in an experimental lab.
The recent trend is for universities to provide central RSE groups (either by core-funding or being costed onto grants in parts) which operate more like consultants. This allows a route in for people who do not have a PhD or a background in a particular field, and can support many smaller projects across the university.
For example at Bristol, we have a combination. We have a lose federation of "embedded" RSEs and some "RSE group" RSEs (funded through various grants) who work together to learn from each other, teach around the uni and provide a support network. This group was able to be formed due to a fellowship from the government back in 2015 (the same fellowship that allowed Paul to start his group at Sheffield).
It's important to remember that anyone who is doing "research software engineering" is an RSE, even if it's not their job title.
A big problem as I see it is that someone not in the first wave of RSEs (who are now mostly in leadership positions across the various research intensive Universities groups) basically gets stuck with no appreciable career growth. You can probably get to a permanent job in the 42-50k salary band as a “Senior RSE”, which seems to be used for people even without that much experience to try and bump up the salary, but beyond that it would be very difficult to jump to the band above without staying in the same role for ~5+ years and hoping that someone left. At that point you’re obviously at top of automatic increment band and don’t get a meaningful pay rise for a potentially number of years. I thought it was very unlikely that there would more jobs at that level because when bidding for time on grants, academics don’t want to pay someone external to their team more than they pay their own post docs, and heard comments to that effect from academics themselves. For centrally funded positions it’s even harder, having seen my old group lead battling to try and get even 32-40k positions approved.
So in my mind, there’s a big problem where someone ambitious could job hop into an easier and better paid job in industry. Why wait a really long time to get promoted when you could leave and get paid more than the top of the next salary band anyway?
An RSE is paid like a standard researcher (with none of the benefits, recognition etc as detailed in other comments), which means that the university will take, minimum 50% of the grant money for that person.
This is for building work, maintenance, general staff, etc etc. But I need a laptop, and I work from home, most of the time the IT staff just get in my way and slow my computer down with the crap they install.
If the universities didn't take that cut, we could easily pay the RSEs more. There's no job security, very little recognition, but every single subject requires good code these days.
The real issue here is that without decent coders our science will fall behind. Any idiot can write bad code, and smart people can write exceptionally bad code, we've all been there, not writing simple readable code because it made us feel clever, till we had to read it a year later. Academics are smart people who believe they can teach themselves to code, and they can, but it's generally not good software.
There are also the ancillary benefits of just having someone around who can do things like rename your hundreds of files by writing a script in 5 minutes rather than you taking two days to do it (this actually happened).
For the sake of western science this should be addressed.
Aside, can I ask why you don't want to work at a college or university?
I think the advice in the interview is pretty sound. OSS is a great way to contribute and become in demand.
But the main issue here is that there are not enough RSE's. "If RSEs can get the recognition and rewards that they deserve, then the career path will be that much more visible and attractive."
Here are my two cents:
1) Academia will never compete in salary with big tech (except maybe in the big Ivy League schools). Universities should promote spin offs, help with patents, run regular hackathons and an incubator program. That will give engineers an incentive to at least start their career in an academic lab.
2) Some of the best science happens in startups and org x labs, not just academia. But even here I found there are very few companies that value an RSE. You're either an RS, or an SE and the extra R or E is in title only. They will see you as a researcher who knows a bit of python, or as an engineer who wasted years in graduate school. You will be interviewed as one or the other. You will be paid as one or the other. I've never seen a company appreciate the value of fully grokking or contributing to the researchers algorithms and Jupiter notebooks and then architecting an efficient implementation at scale. You will also be kept in your swim lane. Researcher? Can't touch the production systems. Engineer? Good luck getting a patent application going. If there is a hiring manager out there reading this... please consider making a true RSE interview and role.
Contrary to many of the posts I see here, it was a great job, except for the below-market pay. I felt like I was contributing to something valuable. It beat working on yet another boring commerce website. I had interesting coworkers and worked with interesting technology.
I didn't experience any of the politics around academia. I felt respected by everyone. I don't have a PhD. I don't even have an undergrad degree. I dropped out of highschool, got a GED, and did a couple years of undergrad, got hired by the university, and never finished. No one ever treated me like I was second-class.
The only reason I'm not still working as a RSE is because I took a 50% pay cut when I took the job and was lured away by money later.
I wonder how many of the negative commenters here have any actual experience as or with RSEs.
Most of my job is about optimizing numeric code, lots of numpy, pandas, numba but also thinking of new algorithms. I get respect and they put my name in the research papers they publish, even if I did not write a word in the paper and have no knowledge at all of the underlying science ;-) Such companies are rare but they do exist. Much better than academia because there is no bureaucracy. I got in because of a friend of a friend...
For example, in Germany where I now live, all public universities (and almost all prestigious schools that do high-profile research are public) have strict salary rules. You can find tables (like this one [1]), according to which your pay is determined. It depends on the job, land (state), degree, and years worked. While this kind of money allows comfortable life (or at least did before the inflation hit hard), industry offers at least 30% more (based on my perception).
I guess science does not need more research software engineers after all!
A major pain point in RSE is building and sustaining larger software teams around a a single, focused long-term project with ~6-8 RSEs working together. Even the most well-funded and successful academic labs have a limit to what how much RSE effort they can support.
We need new funding models where engineering-focused leaders can muster sufficient resources (autonomously) to build great software teams that maintain high quality software in the long run. Focused Research Organizations (FROs) seem like an interesting new idea (https://www.dayoneproject.org/ideas/focused-research-organiz...).
The worst parts of that job are the same for every body in research that isn't a PI pulling in lots of grant money: low pay, terrible or non-existent benefits, bureaucracy ran by non-science people, making do with older equipment.
There is also a breaking in period where you have to prove you are not a moron. Since you come from a different educational area, and most likely aren't as educated, this can take some time and effort.
The best part of the job: getting to work with highly educated people, getting to see science be discovered, being able to actually plan and execute projects because everything isn't on fire, regular business hours with no after hours support calls.
Matthew J. Lambert, Wayne O. Cochran. Kyle G. Olsen, Cynthia D. Cooper,
Evidence for widespread subfunctionalization of splice forms in vertebrate genomes,
{\em Genome Research.} 2015 May; 25(5): 624Ð632.The pay is indeed around 60-70% of what the industry offers, but I would have the same rate if was making a CRUD app[0] for the same large company, so it's more about the employer than the type of work.
[0] which is something I did in the past - a webapp generating documents ensuring regulatory compliance. This organisation likes to make webapps.
Coding somebody else’s problem lacks status in academia, and the money is poor. Why would people do this as a career?
I worked as a research engineer for a few years. They are really needed but unfortunately this will be a waste of time, effort and knowledge. The real problem with software development in science is that it is driven by the scientists who are (mostly) short sighted in their sw dev planning/schedule, they do not appreciate software debt.
And given their budgets and their focus in publishing, I understand their decisions.
I am sure a lot of academic code would be written by the open source community, but alas , researchers are very stingy giving out their data
I looked into scientific programming, and everyone I offered to work for if given an actual position, not an assigned seat and the perpetual threat of firing if a string of soft money grants ran out, and I got zero takers.
The science field is full of performative narcissists who think if they're trying to "cure cancer" or whatever, the lessons from psychology, behavioral economics etc simply do not apply to them.
People wonder why r&d in the us is in a slump... Well it's mostly because choosing a natural science for a profession is punishment compared to software.
Maybe I can steer you towards some Github repos
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6209097/
Not an expert but this looks like the most popular repo:
https://github.com/gromacs/gromacs
Good luck!
Not just in academia, but also in biotech/pharma, where MDs or MBAs are too often the top dog.
This largely explains why Pharma and healthcare is so behind when it comes to tech
(Exceptions are the new, small biotechs. But the larger ones are more like dinosaurs)
If Science wants more research software engineers then they should start treating them like peers rather than hired help.
Everyone has the same thinking that we will be here 40-70years and our finite time is worth more and more.
Im sorry for the ppl tho dont have options to do better for w/e reasons.
This is probably not going to happen unless my employer decides to pay something competitive with industry. I'm here because I want to be. There is an opportunity cost here and not many people are willing to pay it.
Is there room for something similar in academia - to bolster the profile of the RSE idea, to get some support funding etc?
It's a problem with research institutes that they consider themselves academics and look down on people with extensive work experience.