mbrundle

Related course (w video lectures): https://news.ycombinator.com/item?id=9311752

It's sobering to read that they know that WeChat is flawed and yet continue to use it because solutions like Telegram won't be adopted widely enough by the community.

Printastic | Junior Data Scientist | London, UK | Python | ONSITE | Full Time | rehan@getprintastic.com

Printastic is a mobile app for building attractive photo books from a user's photo collection with minimal effort (www.getprintastic.com). We're looking for a junior data scientist to collaborate closely with the core team on building several machine learning / computer vision based projects. You'll need a technical graduate degree (or strong undergraduate degree), demonstrable experience with Python and machine learning, a keen interest in learning the latest developments in ML/CV and an ability to learn new skills quickly (e.g. Swift, TensorFlow). The role would be ideal for someone who is looking to gain industry data science experience.

Interview process is an initial phone screen, a 2hr onsite interview (which involves discussing your previous projects and a recent journal article), and a final phone call with one of our technical advisors.

We're based in Farringdon. If you're interested, please send a CV + covering letter to rehan@getprintastic.com.

I think this is really exciting. A year ago, I was looking for something exactly like this - a formal course in Computer Science, with access to top mentors, which could fill in the theory gaps in my mostly-self-taught background, and which I could do in my own time whilst juggling looking after my young kids. Back then, the options were either MOOCs (free but no consistent mentorship available), remote bootcamps like Hack Reactor (expensive, limited syllabus), or audit a course at a local uni (also expensive in the US, limited access to resources, have to jump through annoying course prereq hoops). Or, of course, commit to a full-time degree, complete with the long-timeline application process, the expense and the relocation hurdles. More recently, several online CS Masters degrees have been made available but at eye-watering prices. (UC Berkeley's Data Science MSc weighs in at over $60k!)

This fills the need perfectly for mid-career types who want an affordable option that fits in alongside their work and family life. I think there's an enormous unmet demand here, particularly in an era where professionals need to keep retraining throughout their careers in order to stay relevant. I really hope that this opens the floodgates and other top institutions will follow suit.

Printastic | Data Scientist | London, UK | Python | ONSITE | Full Time | rehan@getprintastic.com

Printastic is a mobile app for building attractive photo books from a user's photo collection with minimal effort (www.getprintastic.com). We're a small startup going through an exciting rapid growth stage. We're looking for a second data scientist to collaborate closely with the core team on building machine learning / computer vision methods for automatically generating candidate photo albums, predicting user behaviour from their app usage data, and other projects. You'll need a technical graduate degree, demonstrable experience with Python and machine learning, a strong interest in keeping up with the latest developments in ML/CV and an ability to learn new techniques and languages quickly. (CV experience is a bonus but not necessary.)

Interview process is an initial phone screen, a 2hr max onsite (where you'll be asked about your previous work, we'll discuss an academic journal article, and we'll brainstorm a project), and a final phone call with one of our technical advisors.

We're based in Farringdon. If you're interested, please send a CV + covering letter to rehan@getprintastic.com.

Very interesting littlw study. The results seem to make sense, and I'm impressed that his model can learn what causes his weight swings given the low resolution delta-weight data he collects.

The author uses vowpal-wabbit to train his regression model. Anybody know what learning algorithm it uses (eg random forest?) Here's the link: https://github.com/JohnLangford/vowpal_wabbit/wiki

I would ignore the top reply. I found this a very helpful collection, and I appreciated your breaking down of resources for starting off in different areas of deep learning. It's at the perfect level for a summer student that I'm currently supervising.

Here's the link to my post on integrating Matlab and Python, in case it's helpful.

http://www.hacker-dad.com/integrating-matlab-and-python/

I do a lot of image processing algorithm development (e.g. building custom segmentation algorithms). I started out using Matlab during my DPhil research because (a) it came with very fully featured toolboxes, and (b) most other researchers wrote their code in Matlab too.

Recently I've started making heavy use of Python at work. It's made significant strides in terms of its image processing libraries. I haven't used scikit-image much, but we routinely work with OpenCV at our office, because its feature set is a great complement to Matlab's. However there still remains a lot of really useful 3rd party code that's only accessible with Matlab.

Our current workflow, as a result, is to use Matlab and Python where necessary (because different problems are more tractable with one or the other), and to then stitch them together. This is a bit messy but doable. (I wrote a blog post about it on my site if anyone needs to look into this.)

Bottom line - the Python libraries are fantastic for dabbling with image processing, but for serious work, depending on what you're doing, you may still need to get your hands dirty with Matlab for the time being. This is unfortunate, because Python is free and Matlab is $$$. This will likely continue until a large enough proportion of the academic image processing / computer vision Matlab researcher userbase switches over to Python.

I've noticed that brit humour seems to appeal to a particular subset of Americans. Not quite sure how to define the ones that appreciate it most, but regardless, it's great to have them in the brit humour appreciation club. Glad to know that someone else out there gets us crazy people, when the rest of the world doesn't. : )

I know I'm hanging out on the right site when a link like this gets promoted to the top spot. : )

A quick note to Hitchhikers Guide to the Galaxy newcomers, your mileage may vary depending on your background. I read all the books when I was a kid, and absolutely loved them. And when the most recent film came out, a bunch of engineer friends and I went to see it. Me and the other British friend howled with laughter throughout, but the two other (from India and Brunei) looked utterly baffled. So, it may, based on this limited sample, play better to those with a British sense of humour...

Earlier this year, someone from Reddit asked me for advice on classifying some tricky medical image data (CT/MRI/SPECT of healthy vs Parkinson's patients), an area somewhat related to my PhD research. The guy was very knowledgable and had already made a lot of progress. I asked him which lab he was doing his thesis in. He replied to say he was a high school student.

I'm seriously impressed with what these students are capable of these days.

I'm a bit puzzled by this article. Sure, the tech media didn't question it very much. But then again, I never really associated the tech media with any level of serious investigative journalism. I doubt they have the time, manpower or ability to really dig deep behind the PR barrier. Plus, as the article mentions, they make a lot of money from keeping companies happy.

If I want someone to give me the lowdown on what's really going on, then I'll turn to the NYT first, not TechCrunch. Surely I'm not alone in this?

My thoughts exactly. I've never known a journal to do anything to promote the fact that someone has reviewed for them (eg produce a letter on request that states which articles someone has reviewed for them). It'd be a nice and free way of creating value to reviewers who generally work for free, but I don't know any journal who's ever done this.

On paper this looks really interesting, and I really hope it does well. Immunotherapy is one of the most promising recent success stories of cancer therapy. However I remain skeptical of any efforts to get academic researchers to work together, because (a) the pressure to publish-or-perish is so great that I can see them retreating back into their silos, and (b) academic researchers just aren't used to collaborating. The closest they get is to meet up once every few months to discuss each research group's progress. For this to work, they'll have to find new ways to share their progress, which can become part of their working culture.

> ...isn't this just part of being an adult?

It is, but in practice I don't think it's as straightforward as that, because:

1. There can be a presumption that the contract will be fairly written. Some of my recruiting employers have behaved as though they have your best interests at heart, which could lull you into a false sense of security. [A]

2. As you mention, it can seem that the decision is strictly yes/no. Corporate EULAs fall into this category - you either agree to abide to the terms of huge piles of legalese, or you don't use the software. I admit that I just agree to all of these. Even if I did read them and find a clause that I wanted to negotiate, I don't know how I'd even begin to broach the subject with an Apple, Microsoft, Adobe etc. And a similar situation exists in employment contracts, where someone who is desperate to accept a job offer (e.g. at a big uni or company), and they may not know that the terms are negotiable.

3. People just may not be aware that there are really long term consequences, such as the non-compete clause in the OP. I certainly didn't know about these sorts of issues before I worked in the Bay Area.

4. Some of the more restrictive legal clauses may be quite innocuously or confusingly written, and therefore require a lawyer with relevant experience to spot. So even if a layperson is being diligent, they may still fall afoul of something dodgy.

========

[A] Quick aside. One startup CEO gave me such a contract, and when I verbally raised some issues, he just laughed them off and said they were typical boiler-plate legal language, and said that none of this would ever affect me. I was on friendly terms with the CEO and was therefore ready to accept those assurances, but my wife wouldn't have any of it, and she urged me to make some modifications.

I just looked up the email thread where my wife and I discussed one particular startup employment contract. It featured a section on intellectual property rights, where if you didn't complete an additional "Exhibit A", you were at risk of giving the company free licensing rights to anything you'd worked on in the past.

Wife: "It should be okay to sign, but not without completing Exhibit A."

Me: "I think it'll be fine, they're nice people, I'll just go ahead and sign it."

Wife: "Wrong answer."

I spent hours working on the Exhibit A, and took it to the CEO the next day. When he read it, he saw red and stormed off out of the office to talk to his legal counsel. The legal counsel came back and said yeah, my wife was right, and agreed to do the changes.

When I joined the company, it turned out I was the only technical person who had filled out that Exhibit. Would this have affected any of them? Who knows, but there was certainly a quantifiable risk involved. I'm certainly glad that I had my back covered by it when the shit hit the fan. The others were smart people who just didn't understand the legal consequences of not filling in that Exhibit.

I've worked for several startups (US and UK) where the contracts contained very broad restrictions like this non-compete. Fortunately my wife is a lawyer and has helped me to read and red-line them, and then push back and get these sections rewritten. This is something I would NEVER have thought to do by myself, and I don't know many other tech people who do this either. I don't know the legality of these clauses, but my impression is that the companies don't really care - it's easy to put overly restrictive language into their contracts because they have little to lose out of it, as an employee is unlikely to undergo expensive future litigation to sue over it.

Bottom line - there needs to be (a) some low cost support mechanism for people to be able to check whether their contracts contain anything overly broad that could hurt them later, and (b) some way of educating potential employees that they should read contracts in detail, and that they can challenge contracts and push for changes before signing them, and not to treat them as a binary sign as-is/no-sign decision.

Academic journal article reviewing is a very peculiar world, and I wasn't impressed by what I saw of it. My observations (as a former postdoc in biomedical research) were:

Reviewers get no pay or remuneration, pauce guidelines, and no training (which is where articles like this can make a difference). There are no tangible career benefits for doing it (other than 'everyone else does it'), because you won't get any sort of official record for papers that you've reviewed. (And because it's single-blind, you'll never be credited on the paper.) There's very little feedback or quality control on reviews exerted by editors. You can't ever discuss the paper with your fellow reviewer(s). And it's an enormous time sink - reviewing a paper properly takes at least two hours, depending on the length and complexity. This is a real issue when you're in a field where doing lab research, writing your own grants and papers, reading the latest literature to keep up-to-date with the field, and possibly doing some teaching or admin, already takes up most of your time.

It's a seriously broken system. I inherently like the idea of doing reviews because it feels like you're giving back something to the community, but it ended up feeling like this good will was being taken advantage of by the journals, particularly the for-profit ones. I'm amazed that the whole system continues to work as well as it does.

I can relate to the experience of finding everything is totally different now to how it was before you took a break. I was a self-taught developer who then took a lengthy gap to do a PhD and postdoc. That ran into a dead end, and after much reflection, I decided I wanted to get back into coding. My blog has several posts about this thought process which may or may not be helpful. To cut a long story short, I did two online boot camps (one on Ruby on Rails, one on Swift iOS). The experience of coding with an experienced mentor that I could bounce questions off was really helpful, and I soon got that buzz for coding back. I'm now into my second coding job (as an image data scientist for a London fashion tech startup) and I'm loving it.

Please feel free to reach out if you think any of my experience could be relevant and you'd like to chat further.

I've also had issues replicating ODE models from papers in the past. I assume you've tried contacting the authors to ask for clarification and haven't had a reply? If so, I think you should submit all 3, because without any original code available, your effort is likely to be representative of anyone else's. (And it's easily vettable as it'll be on GitHub). The onus is then on the original author to point out how (if at all) your implementation differs from theirs.

This post sums me up too - trained in a very mathematical field, but without a maths background - walked away from it all for a year to rethink things and spend time with my kids - spent ages trying to learn new tech and boost my skills - just got a job again, doing pretty much what I was doing before. And in a month, I've learnt more through my project work as collaborations with team members than I did when I was self-learning during my year off.

Now I'm combining the two - learning some new stuff during my commute, putting it into practice at work - and at the same time, making sure that my evenings and weekends are free to spend with my family. Right now I couldn't be happier - I feel that I'm finally beginning to develop as a software engineer, but I can maintain a nice work-life balance too.

Totally second the other posts here - learning is good but you really need to be working on building something, because that allows you to really consolidate when you've learnt. It's when you spot the perfect application for some theory you learnt in your project work, and you can execute that application, that you really start getting somewhere. IMHO. :)

Sounds good!

Likewise. I'm also mid-30s with 2 kids, and I'm currently pivoting from a science career to a software developer one. I'm currently deciding whether to go down the self-started startup route or get a normal job, and I'm blogging about my experiences at http://www.hacker-dad.com . Combining parenting and startups presents a set of pretty unique challenges, and I'd love to find other parents who are grappling with them, figure out what it takes to overcome them, and then share the results with the community.

Me too, I thought it was a startup called Jacuzzi that had just been kicked out of China.

Thanks for posting this. Given you were dealing with superiors, you treaded a fine line and did so well. I hope this encourages others to be similarly proactive whenever they see anything that appears to be harassment, regardless of whoever is instigating the act.

No you're not. My background is biochemistry (UK degree), biomedical image processing (UK PhD) and web development (self-taught). After a decade of academic research in medical imaging, I'm now moving more towards pure tech. I'd love to go back to uni, say full time for 6-12 months, either in the US or UK, and rigorously study the fundamentals of Computer Science, however I haven't found a mechanism that really works for me.

An additional undergrad degree is too long and slow - I know I can study MUCH more efficiently than I could when I was 18, and I also want to be very focused on CS, so most of a US undergrad course would be a waste of time to me. On the other hand, a 1 year Masters is too advanced, since I don't have a firm grasp of the basics.

Online courses are definitely filling the gap but I feel like I learn faster with human interaction and mentoring. Bootcamps are okay to an extent (I'm currently doing The Firehose Project to learn some industry web dev practices, and they provide great mentoring and group projects), but they only advance your skills so far, and probably not up to the level I expect you'd be at after completing the first year of CS. It'd be great to see some different options out there which provide the rigour and cater to a more mature audience.

I had almost exactly this issue with Skype earlier this year - they locked my account from using credit, meaning that although I can still log in and make free Skype-to-Skype calls, I still have $10 of Skype credit that it prevents me from using (as I got locked out soon after an auto topup). From what I could tell, my account was compromised for a week, and there were several spurious messages being sent from it, until I finally realised and changed the account password. I emailed customer support several times to unlock it, and provided a whole host of legacy data about the account (which had been set up back in 2005) and how it had been used over that time, but none of it helped them to confirm my identity. Out of frustration, I wrote to them to get them to point out exactly where it said that they could take this type action. They couldn't give me a clear answer, and like the above poster, they suggested I just create a new account. Here's the (almost kafkaesque) replies I received from them:

  > Hello,
  >
  > I understand that your concern. Please see all the 
  > answers to the following questions. 
  > 
  > 1. To have this issue escalated to a senior manager.
  >
  > Please be informed that this case has already been 
  > investigated by the higher department and the result 
  > is that the account should be remain restricted.
  >
  > 2. To have it pointed out where in the Terms and 
  > Conditions it says that you're unable to tell me 
  > the underlying reason for the restriction.
  >
  > I am sorry but we cannot disclose the specific reason 
  > why the account is suspended. You may just refer to 
  > Terms of Use number 11. 
  > 
  > 3. To have it pointed out where in the Terms and 
  > Conditions it says that you're unable to refund me 
  > the money on my account. OR to have the $10 credit 
  > on the account refunded to my Paypal account.
  > 
  > Kindly review Terms of Use number 10.6. 
  > 
  > 4. To have all my payment methods removed from the 
  > account, to avoid the potential for abuse.
  >
  > I would like to apologize but we no longer have the
  > option to make changes on the account since it has
  > already been suspended due to this kind of concern.
  >
  >5. To have the account completely shut down, not just 
  > restricted. (OR to have it pointed out in where in 
  > the Terms and Conditions it says you're unable to do this.)
  >
  > It is already our internal rule that we cannot make 
  > any changes on the account like cancelling the 
  > account once it has been suspended or restricted. 
  >
  > I hope that I was able to answer all your inquiry and 
  > I would like to apologize if we can no longer assist 
  > you further regarding your concern.
  >
  > Best regards,
  >
  > Diana B.
  > Skype Customer Service

====

  > Hello,
  > 
  > I am sorry to know that your account has been restricted. 
  > I understand how important it is to recover your account.
  > However, your account has been detected that there are 
  > some activities which are contrary to Skype's Terms and
  > Conditions have taken place via your Skype account.
  > 
  > Following our investigation, we are unable to reinstate
  > your account. Please understand that we are unable to
  > explain the specific reason for restricting your account
  > and this account needs to be remain restricted.
  >
  > However, you are always welcome to create new account
  > to continuously enjoy the free services and features
  > from Skype.
  >
  > Best regards,
  >
  > Diana B.
  > Skype Customer Service

To add insult to injury, I notified Paypal (which I used for the auto topup payments) to reverse the charge, owing to misconduct by Microsoft in refusing to unlock the account. After a week in which they did "their own investigation", where they didn't even contact me for any corroborary info, they refused to issue a refund. All I got was this 2-line reply:

  > We have concluded our investigation into this case. Unfortunately, at this
  > time we are unable to decide this claim in your favor.

Made me think that they're secretly in cahoots with Microsoft. I was pretty livid about the whole experience, and it's seriously damaged my trust in Microsoft services.

I'm dead interested in this (just applied) and I'm definitely not financially settled - actually the opposite, finished a postdoc and am currently bootstrapping a startup. But the value of this looks amazing, so if I can put the funds together, I'm sure it'll be well worth it.

I just tried buying an AutoRip-eligible CD and it treated the AutoRip mp3s just like a regular mp3 purchase - I was able to download them using the Amazon MP3 Downloader app.

Wonder if Amazon will soon lower the prices of MP3 albums which are pricer than the CDs to match the CDs, or if they're somehow prevented from doing so.

I doubt they get much say over what goes into their employment contracts. Added to which, you assume all potential employees are savvy about these issues, which seems like a stretch. I can think of several techies who blank out at the mention of IP and wouldn't appreciate (or care less about) the consequences of such clauses. Particularly not when there's a lucrative dream job at a big firm at stake.

If you're going to name-and-shame anyone involved in filing a patent (and I'm still doubtful about what this would achieve, given all they're doing is maximising a flawed system), the lawyers would make more sense (you can find out who they are by examining the filings on the USPTO PAIR database), or the upper management who determines patent policy.