American Employers Are Hung Up on Hiring PhDs
bloomberg.com
bloomberg.com
I've heard and observed that the Ph.D teaches you one crucial skill - how to take a vaguely defined research problem that nobody has done before, where you're not even sure that a solution exists, and make tangible, rigorous progress on it. Your actual doctoral topic as a Ph.D student will probably be useless. Most of your coursework is the same as what any master's degree gives you. But the experience that sets you apart from B.S. and M.S. graduates is that of slogging through a dissertation, trying to make an original contribution to knowledge.
There are other ways to develop this skillset. Founding a company does it as well, as does developing a new product from scratch. High-level creative work (eg. writing a novel, getting traction for your band, producing a theatrical performance or radio show) does as well. Companies love to hire candidates from these backgrounds as well, but most don't want to work for someone else. (Interestingly, depression and anxiety might actually be a feature of developing this skillset, not a bug - these two mental disorders are endemic among these fields as well.)
The reason employers get hung up on this is because the economic returns to innovation are at an all-time high these days, and so developing that blockbuster new product or highly-efficient new process makes a lot of money. And part of the reason returns to innovation are so high is because so few people are actually innovating. While there certainly exist people with just a bachelors or even no degree at all who are capable of innovating, a large number of bachelor's holders do not have the toolset or mental fortitude to break out beyond what everybody else knows and create new knowledge.
I like the way you put this. I am researcher in academia, and my parents (who do not have advanced degrees and have never done research) always ask me "When you get into the lab, how do you know what to do?".
No one really tells me. My PI only gives very general research directions. It's up to experience and intuition, part of which was gained during my grad student years.
But also agree there are other ways to learn that as well. There will never be a one-size-fits-all solution.
Of course you can learn things without this structure, but having a phd is a reasonable short hand for ability in independent innovation, perhaps companies lean on it a bit much, but that is not unique to phds.
The main problem I've had bridging phd's to commercial work is hitting the right focus on skill development in areas that were deficient/missing in their course of study. A classic example, a lot of technical phds who have programmed every day for many years ... are not skilled programmers. They can become so, but have to want to learn (and not dismiss it as "trivial")
I had never thought about degrees in those terms, but that feels like a surprisingly terse and accurate description. Thanks!
When I've worked with non-PhDs in a science environment, it's the rigorous breakdown of experimental approach that's often missing. Unless the individual had a really strong track record demonstrating this skill, I'd heavily favor PhDs for biotech/science positions on a leadership track.
That said, I'm quite biased, have to justify the years I lost somehow :)
I dare you to write a study to excavate and delineate the potential effects of this bias on hiring.
Moreover, I only have a bachelor's, but I had 6 years of training and humiliating embarrassments from those 30% of people, which I believe strongly contributed to my experimental rigor. I dont know that every person with a PhD has dealt with that kind of pain and would be biased against people that haven't dealt with that pain.
I'm sure you are familiar with the concept of fuckyou money. Its a different number for different people.
Say that number is 5. I will give you many examples below. Every example is a real living person. I personally know every one of these people.
(enterprise programmer in stodgy bank in usa) 1 + 1 + 1 + 1 + 1 = 5
(enterprise programmer in stodgy bank in poor country) 0.5 + 0.5 + 0.5 + ... + 0.5 = 5
( steady progress in startup until faang acquired it) 0.5 + 0.8 + 1.2 + 1.2 + 1.3 + 6 > 5
( many failed startups in valley until 1 success) 0.5 + 0.8 + -2 + 0.5 + 0.7 + -2 + 0.5 + 20 >> 5
( phd student whose invention was purchased by salesforce ) 0.1 + 0.1 + 0.1 + 0.1 + 0.1 + 0.1 + 15 >> 5
( phd student who ended up in fang ) 0.1 + 0.1 + 0.1 + 0.1 + 0.1 + 0.1 + 2 + 3 > 5
etc.
- I made more on my first day of work than almost any of my professors will make on their last day work.
- I also personally know 3 PhDs at tech companies in the Bay that make substantially less than me despite being 4-6 years older than me.
- I am also acquainted with two people that made an obscene amount of money from their tech skills; one maxed out at a high diploma and the other a PhD.
- The other 20 or so PhDs I know don't make substantially more than I do.
- I've also seen a few PhDs at my last company end up as systems engineers which is incredibly depressing because they could have been employed in that same role if they stopped at a bachelors.
It’s only depressing if they feel like they didn’t get anything out of their PhD :)
I suppose it might not work out in such a way for all research though. I was lucky enough that my coursework left me with enough math and statistical skills to either learn or fake my way through most relevant math intensive computing stuff while my actual work was things like parallel optimization schemes for neural network training (and lots of programming in the big frameworks)
I’d also say that 30% is still not a bad ratio and better than most training sources on average.
Still it’s not a perfect screen by any means.
In undergrad I had a really nice meta course called GEOG 100: On Becoming a Geographer. Loved it. One slide I remember was that a bachelor's was becoming acquainted with the discipline. A master's is to become an expert in a domain. And a PhD is to drill deep into a domain and uncover something new.
Breadth vs. depth basically.
I'm not convinced, or I should say my experience does not support this. I think it is more likely that those that already have that skill are the ones that pursue the Ph.D path.
I don't think all Ph.D. students or any ordinary engineer can take a vaguely defined research problem and make it succeed. Ph.D. doesn't necessarily prepare you for the research part; it does, however, prepare you to advertise a very /bad solution/ as a novel contribution. It wasn't always like this, but it has come to it. It takes credibility, curiosity, character, and of course, research skills to solve a vaguely defined problem---none of which are given to you by a Ph.D. degree.
I have worked at two FAANGs, and more often than not my interactions with Ph.D. degree holders have left a bad taste in my mouth. Speaking of which, there was one person that was "selling" an event timeline as a root cause analysis system that does "temporal" correlation (with no filtering or association at all) :). And another person that was advertising a DFS compilation of a neural net during the training phase (as opposed to the typical BFS that people do) as a superior and novel contribution that changes how we think about neural nets or something along those lines.
A good engineer would have laughed at both after carefully considering all aspects of the problem.
I suspect that Ph.D. "engineers" are more desirable because of their broader skillset (they have worked with more tools and have taken more classes) and also the fact that companies can hire them at almost the same cost as a BS/MS degree holders. Plus universities have already done some filtering on Ph.Ds.
Some PhD-holders were self-motivated before. Others learned it at school. Still others were motivated, but unfocused, and had to become disciplined.
Everyone who has a strong opinion on the topic is, by definition, extrapolating from limited data. Moreover, even if every single PhD learned nothing in school, it’s still a useful signal if it enriches the pool of candidates who are self-motivated and independent.
The first is worth their weight in gold, the second is fairly easy to detect after a few git check-ins.
Actually I lied, there's a third kind, brilliant people that get bored with their job and write super-complicated frameworks to satiate that boredom. That's not restricted to PhDs but I see it more often with them.
Reasons might include: couldn't make it in real jobs and delayed joining the the job force for as long as possible
doesn't want to be held accountable for actual working solutions and likes working towards "novelty"
enjoys approaching every problem with the most complicated possible solution they are aware of
thinks generalization is the only way to approach specific special-case problems, if they can't find a generalized solution then they think its unsolvable even when the special-case solution is fine
has grand research ideas they weren't able to convince anybody in academia to fund and think they can chip away at those ideas by cowing coworkers into working on it for them
likes being the "smart guy in the room" and builds a career of talking a lot with complex technobabble and buzzwords but has no idea what they're saying (I've encountered many many of these myself) -- often has mile long CVs as well
creditialists who think racking up pretend education credits is the most important thing in life
failed academics who point to their large private sector salaries as why they've ended up on the superior path than their old peers
and so on....these are literally descriptions of people I've encountered just in my current job over the past 5 years...it keeps going on and on.
Considering a PhD is supposed to provide skills for learning and tackling problems, the incredibly high percentage of PhDs I've worked with that are completely unable to do so has really really put me off of the credential.
Few better ways to take that to the extreme in many cases than a PhD.
Also, as a foreigner, can't help but point out that there seems to universally be a lot of assumptions made with academic qualifications and the like with many countries immigration laws and systems...
You might have a bit of that skill to start with but a PhD lets you practice and hone that skill, which you wouldn't be doing in a standard coding position.
Given how insightful your comment was, I was wondering if you could elaborate on this. Why, outside of comorbidity, do you think depression and anxiety might be related to developing this skillset?
Today, when you're tasked with creating something new, and your identity is invested in such a mission, it's a much more gargantuan task than being indispensable to your local tribe. Competing at a global level is just not something we're generally wired to do. I would be curious if world-class athletes, movie stars, business executives, politicians, etc. also have similar emotional problems.
But our emotions do not work on such a scale. They're meant to deal with the anxieties of what a few hundred people in your community think and how to secure your next meal.
In creative lines of work (however one chooses to define it), one is trying to be the superlative best or first to achieve something valuable on the scale of billions of people. And you probably hope that such a creation outlives you. We have not been designed for such tasks at an emotional level.
But what I've noticed is that aside from frequently co-occurring with creative work, depression and anxiety are frequently the trigger to finding a non-obvious solution and making forward progress. In other words, they're not just mental disorders; they're feelings that can give you useful signal about how to proceed next, if you listen very carefully to them. Anxiety is basically your emotional systems being hyper-engaged, such that you get an emotional response from nearly everything. Depression feels like the opposite, a deadening of your emotional senses, but oftentimes escaping the depression actually points the way to escaping your real-world problem. If you can imagine a future where you are not depressed (which if you've suffered clinical depression is actually quite difficult), very likely that's a strong signal that that's what you should actually be doing, no matter how weird or unreasonable it sounds.
There's some evidence for this hypothesis here (2009):
https://www.scientificamerican.com/article/depressions-evolu...
You could also think of the anxiety related to creative work in relation to System 1 of the brain in Daniel Kahneman's Thinking Fast and Slow. In highly ambiguous circumstances (which basically all creative work is), it makes sense for your brain to become hypersensitive to all available stimuli, and to quickly pattern-match on a course of action without too much rational consideration. We experience that sort of emotional overload as anxiety.
Interesting, that was not the impression I had from 2 engineering professors at my local uni.
It sounded like I was going to be their assistant in studying their idea for up to 6 years.
Pretty accurate. I was asked in a job interview once "what practical applications does your research have that would benefit us?" - they were surprised when I said my research results weren't useful to them at all, but the skills and broader knowledge I gained doing the research would be.
It's very uncommon you go seeking a company where you can directly apply what you did your thesis on. Your thesis topic pretty much has to be niche and obscure if it's on the edge of current knowledge.
I’d like to elaborate on this a bit further. In my experience as an academic researcher in a technical field (math), one of the most important skills is the ability formulate precise statements/hypotheses that satisfy several criteria. First, they should relate to the existing body of research in your field in a meaningful way. Second, they should have non-trivial answers (in the context of math, i mean that the proof of your hypothesized theorem should be legitimately nontrivial, not merely obfuscated to appear that way, as good mathematicians will be able to tell) or reveal something that was previously unexpected (such as a connection between two previously different topics). Finally, they must be testable/verifiable/provable/etc using your available methodologies and within a predictable timeframe.
My dearest mentor, a man who has forgotten more about *nix, coding, and about anything else than I will ever know has a HS diploma and can and does run rings around everyone else. He has a mind like a steel trap. All of the great coders I have known had, at best, a BS or BA, none of them in an actual IT discipline.
Anyone can learn to code, be a great sysadmin, or a network guru. All it takes is want power. If you want it bad enough, you will get there. In fact, I'll hire a person who has changed careers into IT who is smart and hungry over a simpering PhD candidate who thinks they are the cat's pajamas. There are some calls for PhDs, but not many. It's a badge of some sort for companies to show they have "smart" people. PhDs are great at research, not so much at actual doing things. This has been my experience with working with more than a few.
I definitely could got where I am without the PhD. Hell I wouldn't have needed the engineering degree either and I'd be pretty rich by now, but I don't regret doing the degree because I learnt loads and had fun doing uni things.
The work is totally different; they're not substitutes.
When some people FB Core Data Science came to campus, they made it pretty clear that they were looking for people with doctoral training for their research work (and not, for example, a Master's graduate who completed PhD-level coursework), so I guess it depends on the task/team.
If you have a stack of resumes you're likely to interview everyone with a PhD, but only the ones with the most stellar github profile if they have no formal education.
During the interview if the PhD says something silly you'll probably disregard it. But the person without a formal education needs to be on point.
These and other filters means that a mediocre PhD can get through the hiring gauntlet but only the most kick ass high school graduates can make it through.
My mentor long ago likened these people to Eloi and the IT grunts to Morlocks. I cannot say I disagree.
Everyone is the hero of their own story. It sounds like your mentor is arrogant, and could use more intellectual humility.
>By the year AD 802,701, humanity has evolved into two separate species: the Eloi and the Morlocks. The Eloi live a banal life of ease on the surface of the earth while the Morlocks live underground, tending machinery and providing food, clothing, and inventory for the Eloi.
Please explain how is this arrogant?
Is the condescension and self-importance more clear now? It's an Us versus Them mentality which patronizes other people and looks down on their work.
In my experience, high performance in difficult degrees at good schools are a very strong indicator of future success.
We find gems outside the above profile, but it holds true even across the people we've hired from code schools.
I don't work for a FAANG.
50 years ago big companies would take you with just a HS degree, and if you were smart and showed initiative, put you under the wing of an engineer - or pay to send you to school. This does not happen anymore.
You're also SEVERELY limiting your career if all you complete is high school - you might be OK technically for now, but you will lack all kinds of foundations that would allow you to stay technical, and you'll certainly be disregarded completely as you age for any kind of leadership track.
I think the root of this myth (that having only a HS degree you can still get a technical job) is twofold: One, a currently hot hiring market, and two, the "be what you dream!" movement that was big in the 90's.
If you've worked through a recession (or even just a dulled market), and seen people around you lose their work, their homes, heard their struggles about how hard it is to find work - people would stop parroting this mantra about having minimal education. It's really doing damage to people who hear it.
I've been getting paid to build web apps since 2000, got in right when the first dot-com bust went down; I have never been out of work since entering this field.
I also tend to work for smaller companies.
I also don't regret dropping out of the program before finishing my PhD. After I finished my Master's and started pushing towards the PhD, I came to the realization that what I really wanted to do was just to build cool things, not spend my life in the publish-or-perish cycle of academia. So I stopped with my Master's, took a full-time staff position doing more-or-less exactly the same thing I was doing in my Research Assistantship except for a lot more money, and moved on with my life. And eventually got recruited to where I am now.
> Anyone can learn to code, be a great sysadmin, or a network guru. All it takes is want power. If you want it bad enough, you will get there.
For a second there I thought I was reading some motivational speaker material. I'm not even going to try to explain how wrong that is.
But you're not going to build the first self-driving car without an army including lots of PhDs. You're not going to develop a new pharmaceutical without an army including lots of PhDs. You're not going to get the sort of branding that being the first to beat the best humans at Go/Chess/Poker can get you without an army including lots of PhDs. And so on.
You're absolutely right that PhDs aren't required for the sort of sys admin and coding work that's accessible to bright high schoolers. But there are very few competent CS PhDs who would even bother applying for those jobs.
Also, your comment isn't germane to the article because the article isn't talking about sys admin and coding jobs. I've literally never seen a job posting for a sys admin job that asks for a PhD. Have you?
Why do you think research doesn't count as actually doing something?
You can't just throw bunch of great coders with high school diplomas at these problems and hope they would workout solutions. Solving any of above problems at the level we solve now would be considered pure magic and utterly surprising just a decade ago. These problems requires extensive mathematical modeling and leaning on research with incremental progress that has been collectively made by 100s of brilliant researchers over multiple decades. This is why PhD is desirable for "new tech".
In the case of many PhD degrees, you devote yourself to studying a niche topic in great depth for several years. In the end, employers might see that you wrote a paper like "Electron vortex beams with high quanta of orbital angular momentum" (picking some academic paper at random), and have a tough time connecting that to their day-to-day needs. So I understand why employers have been cautious.
Not to mention, a non-negligible number of PhDs are incompetent at doing anything hands-on. And even when it comes to research, many are experts at coming up with overcomplicated solutions to non-existent problems.
I've had this as well along with comments like "academics are bad at X", "academics tend to write code in way Y", "academics can only write prototypes and not production code", "I'm self-taught and never saw the point of University because..." etc.
Generalisations like this are really dumb.
I wish that, just because I can do one type of thing well, people wouldn't jump to the conclusion that I can't also do another type of thing well. As in: Just because I understand math, please don't jump to the conclusion that I don't understand databases and can't write clean code. Just because I can see the "right way" to solve a problem, doesn't mean I can't also appreciate arguments in favor of the cost-efficient way of solving 20% of a problem if it gives you 80% of the benefit. etc. etc.
e.g. "I didn't see the point in going to university because it's all theory and I think practice is more important", "those who can't do, teach".
> I wish that, just because I can do one type of thing well, people wouldn't jump to the conclusion that I can't also do another type of thing well. As in: Just because I understand math, please don't jump to the conclusion that I don't understand databases and can't write clean code. Just because I can see the "right way" to solve a problem, doesn't mean I can't also appreciate arguments in favor of the cost-efficient way of solving 20% of a problem if it gives you 80% of the benefit. etc. etc.
Yeah, it's a combination of insecurity, cliched thinking and gatekeeping. It's quite feasible to get to expert level at many things.
It depends on your PhD as well. Some involve no programming, some just quick prototypes, some more complex projects etc. Generalising how all PhD students must be is stupid.
But if you're not on one of these research-oriented teams, then I think it's easy to look at PhDs on your own team and think of them as worthless when in fact they were trained for a pretty different set of things. There's the thing about judging a fish's ability to climb a tree. People seem relatively eager (see other comments) to rip into people with doctoral training for some reason.
In other places it makes less sense. And it's good not to make generalizations. If we take ML as an example, there are many excellent people without a PhD and also many PhDs that are great engineers and can write code as good as the best engineers.
I meant the teams in those companies as opposed to the companies more broadly (e.g., Core Data Science at Facebook, not Facebook in general). I mention those companies together because they're well-known for investing a lot in research (e.g., by hiring PhDs). And in these cases, they're hiring PhDs for reasons that are totally different from the reasons for which they hire engineers (who may also have doctorates). For example, there is indeed a difference between the institution-level goals of Facebook and Microsoft Research, but that difference is less substantial between researchers at Core Data Science at Facebook and researchers on the Computational Social Science team at Microsoft Research.
I'm making the point that there is a difference in the value of a PhD depending on where in the company you work. For the research-oriented teams, the value of a PhD lies in the fact that you've ostensibly been trained to contribute to what we know, rather than just applying it.
Going along with your ML example, the difference would be like comparing Athey, Tibshirani, and Wager's work on generalizing random forests against building a random forest using scikit-learn. I'm not saying that someone without a PhD can't write the paper that they did, but it's for sure not at all just a matter of who's better at writing code.
Also, anyone thinking about a PhD owes it to themselves to understand firsthand what that means before getting started. There is an excellent, low-cost way to do this by joining a research group in your field of interest at your undergrad university.
In every case, the students I saw who had the roughest time during their PhD were those who had not even bothered to try doing research during their bachelors.
A PhD degree is about conceiving and completing original research. If you've never done it, it's easy to overlook these points:
1. Your research project will isolate you. Almost nobody will really understand your project and few will care.
2. It will be thankless, praiseless work for the most part. Don't look to your advisor for high-fives.
3. Your project may fail after years of effort.
4. Your research group will have a lot to do with your success. You won't even know what to look for until you've seen a research group in action firsthand.
If any of these points bother you now, seek answers to your questions pronto - and don't even think of enrolling in a PhD program until you're ok with 1-4.
To some extent, this is understandable: a typical undergrad CS curriculum is terrible preparation for ML research. I'd even say almost any analytical field: statistics, signal processing, physics, math, etc. is better preparation for working in ML than a conventional undergrad CS curriculum. But PhD in CS with specialization in ML/AI would be ideal for most employers, since exposure to more advanced stats, linear algebra, estimation theory, etc. becomes likely.
The other factor is that ML research is moving quickly and knowledge is being rapidly disseminated in academic conferences. A PhD has an advantage there over a typical undergrad, since they've already spent years learning to parse academic literature.
I have worked at several AI/ML firms. Only in extremely rare circumstances do your problems require PhD level ML knowledge. But the amount of tech debt produced by 'scientists' is horrendous.
Can someone who studied CS 15 years ago and has since worked as a webdev produce robust ML systems? Probably not without significant training. But someone who has done a CS Masters in the last 5-7 years and has shown interest in the subject will run circles around a PhD who never had to write non-academic code.
> But the amount of tech debt produced by 'scientists' is horrendous.
Absolutely. I've been trying to exit a research-oriented position for the last few months because I want to work with competent engineers again.
Also PhD's simply have more years experience in writing code than MS graduates. So your comment about circles don't make sense. What makes you say PhD's write worse code? And please don't compare the code they write alongside their research to production code, because research code is throwaway code (seriously).
Then again, we're not Google Brain, and we're not attempting to do original research or publish papers. We have very concrete goals for systems we want to build, and so far we've been able to achieve success using already discovered methods. Almost everything we do is traditional engineering to build training datasets and deploy models to production on live data. We also spend a considerable amount of time preparing and delivering presentations on our results and services, so believe it or not interpersonal and presentation skills are actually quite important.
1) Have some aptitude for intellectual activity (in a quantitative field)
2) Had to drive a project and face failure
3) Can learn things on their own
Technical expertise is I think usually a distant fourth, since if you have the first 3 you can generally learn whatever is required.
Do you need a PhD to have these things? Definitely not, but it's a simpler weed out process.
But we have folks with deep technical know how and PHD in some esoteric subject of proton-proton collision inside a nuclear reactor core developing ball attach processes on BGA chip packages or working on chip shooters. It makes no sense.
Also, these PhDs are terribly unhappy and usually on H1-B visa locked in and unable to change jobs.
Why did they hire him then?
Using a library is something you can learn over the course of a vacation by working through a book or tutorial. But understanding what happens inside the library is something that you will never just pick up on the job. You do need actual time, actual training, and interaction with people who already understand this sort of thing very well over the course of many months or years.
I would estimate that I have worked with around 30 PhD-holders in my career. Consistent issues: no ability to hit deadlines, no sense of the big picture, easily distracted by irrelevant details.
This kinda says it all.
This signaling game is really wasteful though. As others rightly point out, people getting these degrees aren't really using PhD skills on the job. They're wasting some of their most productive years in a zero-sum status game.
More and more lately, I've come to believe that tech companies should just decline to play this silly game. Hire promising people right out of high school and tech support them the necessary skills --- bring back the E1 level. Issue your own aptitude tests if you want. You'll be at a competitive advantage relative to the people who demand that candidates do well on the credential treadmill.
For example, the article's graph of "PhD Proliferation" shows a 20k increase of PhDs from 1970s-2010s.... but that is a fairly unremarkable presentation to make on its own when the US population has increased 100M people from 1970 to 2010: http://www.worldometers.info/world-population/us-population/
How is the graduate education population different from 1970s to 2010s? What percentage of college students go on to get PhDs? Is the PhD per capita lower or higher today than it was back then? What is the PhD breakdown by field over time? What other factors affect this?
Also, my hypothesis is that employers hire PhDs for their particular research experience and thesis topic relevant to their business; not just because someone has a PhD in a technical field. For example, a hedge fund is probably more likely to hire a PhD quant analyst with a research speciality in econometrics vs. autonomous driving. Therefore, not all PhDs are equal when it comes to hiring because the demand varies.
Further, the author says "Master’s or even bachelor’s degree holders are often highly talented, and many can learn Ph.D.-level research skills on the job as they go." While it may be true, it's not the point. PhDs are hired not just for research, but also analysis and insight. Why would an employer pay a PhD more money than a pure researcher unless they are getting something else out of it? Nothing says someone needs a PhD to be able to research, but there's a difference between knowing a skill and having the capacity to learn a skill over time, which is the distinction in hiring.
There aren't many degrees that would indicate the ability to design and run a good experiment with little guidance, and doing well in a statistical methods class doesn't cut it, sadly.
From my perspective, the best data scientists strike a balance between technical and business knowledge. And it's the business knowledge that PhDs coming straight from academia often lack.
I say “interview carefully” to avoid generalizing, as there certainly can be PhDs that are excellent. It would be wrong however to assume you can glaze over any lack of experience as if a PhD is some special sign of brilliance; frankly, they may suck.
I have seen “PhDs” flounder worse than any new grad, and have basic problems with critical skills like programming. It is frustrating, as you see the slow-motion failures generated when people are given lots of control without really having spent a significant amount of time working in teams or building skills, etc. or even communicating.
1) you need specific, state of the art knowledge, and this person was actively publishing in that precise area with significant impact.
2) you need good general problem solvers with some technical depth in an area.
If you hire #1, you should build a program around them to capture the knowledge. This is probably the most efficient way to absorb this as institutional knowledge. The ratio of these people to other engineering roles should be 1:fairly big number.
If you hire #2, you should already understand that you are not hiring general tech employees, and define roles accordingly. If they are new to industry they will probably have significant skill deficits (as well as significant skills) and you need a plan to deal with it. They probably are strongly motivated by spending quite a bit of their time on innovation. Plausible exceptions for people whose academic work was long ago.
If you're hiring because you think "smart programmer" you're probably doing it wrong, and will probably both regret it, in my experience.
There is a third model which is essentially building a internal academic environment, but that is expensive and long term enough not to apply to most situations.
I would rephrase it as spending time on research. Which might or might not be innovative.
I know for a fact that there is a plethora of applied ML papers where people simply toy around with the sample set and publish their results. There is nothing innovative about it. Although it requires some knowledge in feature analysis and some other in related algorithms and in basic statistics but that's nothing that cannot be picked up by taking a few months of courses in applied machine learning.
So really, in fact the only case where that extra knowledge is justified is your case #1. And even then I am not convinced that these PhDs actually have ever done anything innovative. There is just too much of a chance that they simply iterated on existing research ideas in their field and they can only continue to do so. Which is again not very innovative I think.
Else, Is this a signal merely because "everyone" is trying to pursue grad school to stay competitive with bachelor, and a new benchmark is needed?
In my view it's inevitable that as one moves up the educational ladder, the amount of variation between candidates is going to increase, as will the presence of alternative paths. There is no "hiring PhDs" as if we're a fungible commodity. Much as it would gratify me for people to believe that all PhDs are magical wizards, I can't defend my entire cohort, nor should I want to. My job is to sell myself on my own merits, that's all. Don't hire anybody without looking under the hood, and without considering alternatives to what you think are your selection criteria. That's just good business sense.
At best we're going to reach the same conclusion that was discussed in recent threads about college degrees: The wisdom of economists and other experts disagrees with the collective wisdom of the market on the value of higher education. At worst, we will assign a derogatory label ("signaling") to that difference without understanding it.
PhD means you've had and accomplished that formalized level of training. There are those who formalized training works very well for, there are also those who it's just not a good fit. That's fine. Just like there's a varying standard and quality of people around in this world, there's a varying standard and quality of people in PhD programs. Just by having a PhD degree simply states that you've gone through the program and learn how to do research. There's a major difference between how academia approaches a problem and how the industry approaches a problem, but what I've understood is that PhDs (in engineering) know how to expand and innovate in a much easier and more structured way than those in industry.
When generalized, most research I find in the industry is mostly "look I changed this thing and it's improved accuracy this bit more in my use scenario!" It's a bit less common to find papers or research done (Engineering-wise) in the industry that fully understands the benefits and limitations of the method improved upon, instead it's mostly "see, here I did it and now try it on your own and see if it works for you!".
All tools and methods can be improved upon or all models are inherently incorrect. I want to know the proper "surface area" that the method can be applied to and what the proper parameters are. Too many times (especially with some international papers) I've seen methods applied to problems that had to be "shimmied in" to make it work. It's like using pastebin.com as a CDN to store your ASCII data of a 3D model for your game (seriously, I've seen this happen...). I mean, it's not the right tool, that's not how it's supposed to be used, but I guess ok?
Source: I'm getting my PhD in engineering in Massachusetts. I've also co-founded a startup and am currently going through another similar experience right now with a "project" that we'll also be transitioning into a product. PhD is something I wanted after seeing what's in the industry.
Why do you think that is?
I know very few PhDs who are also sysadmins. And fewer yet who can properly diagnose issues with workstations, servers, etc. In college campuses, the labs are maintained by sysadmins. I've worked with PhDs who were "coders", yet could not set up their own programming environments because they were done for them in academia. Not all, but most. I'm not deriding PhDs here, but as one poster said, they generally work in silo environments and don't really understand the broader pictures around them.
And it takes time to learn how to use software tools effectively in a standard way in a production environment that others can understand and extend and maintain. This isn't taught in academia either, and usually it takes several years in the real world for these skills to mature and integrate into the practice of writing good pro code.
A PhD offers no advantage in this school of hard knocks.
"Oh, we see you have Perl, awk, sed, Bash on your CV.. We need you to write some code to get this and that data from the server logs and then spit it out in this and that format, but we don't want this particular info."
I about crapped myself the first day on the job. Fortunately, my team lead knew what he had asked for and coached me through how to be a useful nix admin. This led me to being more interested in coding than setting up nix boxes and getting them to talk to databases. I fell in love with regex stuff and still love it. Perl and Python are still my go-tos with regards to this. I still write small (25-30 line) programs in awk. Sadly, I have a job now where they all love PowerShell, so I've been "forced" to use it for certain Windows tooling tasks. Learning Go on my own now. I find R and F# fascinating and might take this on in time.
I've often had the task of guiding phd's in making the transition. In my experience, the most motivated ones can do it in six months, but most take a year or two (or never get there).
ymmv, of course.
Secondly, in many cases bright kids from India/China/Korea come to USA to only do a Phd and end up on these jobs giving wrong signals that Phds are important.
As an example, I ask a guy to implement me a LinkedList during the interview. The result usually works, but designwise... Oh, my god...
And that makes a perfect sense: once an article is published you usually don't need to maintain your code for another decade.
While they might be good at research, on the other side they have some limitations when coming to concrete work. The last one wanted just code and skip all the theoretical work (reading docs, specifications etc), bringing us to let him go, because he wanted only his IDE and no more (not even considering other tasks such as taking care of his own work)
The people who were smart in high school are still smart today, regardless of their credentials. Some of the smartest people I know have PhDs, and some do not.
Edit: Or startup pitching, M&A or even some sales - basically our deck team slide is "X PhDs, Y ex-Googlers".
You should spend several years of investing starting with index funds then moving up to more complex instruments before you invest in exotica like hedge funds.
I've seen and worked with non stat major PhDs handling model process especially data. I think master in statistic is good enough for modeling and handling data and more so than others for most traditional models.
For the last 15 years or so I've worked in environments with large percentages of PhD co-workers and have come away with this conclusion. Having a PhD seems to indicate that your brain was stretched to one side or another of extreme human thought and left very little in the middle.
On the one side, you've lost the ability to make concrete contributions to almost any project and are nearly incapable of delivering anything with any focus or impact. Junior staff significantly outperform you in almost every task and you seem to think the workplace is where you can come to pontificate and consider all the rest of the humans here are grad students you can badger around to do pointless and off-focus "research" that's really designed to get lots of papers written for you to take to conferences.
On the other, you can figure out how to perform pretty much any of the work of your peers, but can keep climbing the complexity ladder into abstraction heaven and then often bring fire down from the gods to push intractable problems forward.
I've never encountered somebody with a PhD who fit both categories and at times I've been caught in confluences of multiple people of the first category who made work life so unbearable I started having nightmares about it before quitting. People of the second category are rare enough that its often just more useful to hire people without PhDs if you don't have incredibly hard research problems to work on.
The problem is so prevalent that I've known several people drop out of PhD programs because they're afraid of ending up in category one.
One other anecdote, I've often noticed that managers with PhDs will have an incredibly huge blindspot for people in the first category and allow them to run roughshod over the workplace while penalizing staff with fewer credentials who nevertheless outperform their better credentialed coworkers in nearly every work task. It morale destroying and no amount of discussion or other techniques ever seems to overcome that blindspot. I've had people quit over more senior managers continuing to prop up obviously failed employees solely because they have "PhD" next to their name.
I've never encountered this phenomenon among managers without PhDs.
I'm generally incredibly unimpressed by a PhD on a resume. I care much more for a history of solid performance and delivery. But I'm definitely in the minority and have seen dozens of terrible candidates hired because of those three magic letters.
I'm so bothered by both the academic process of getting a PhD as well as the "product" I've encountered in the real world that I decided years ago that I'd rather pursue another BS or MS instead of a PhD.
this comment nails it far more succinctly than I have https://news.ycombinator.com/item?id=19506719