Want to pivot career to ML, want to stay at Google too

Google SrNa64
Aug 9 18 Comments

Currently a full stack eng, want to work in Brain/research on something good for the world. I also want to stay at G for the foreseeable future.

Has anyone done this? How was your experience? What path should I take? Should I go back to school?

L4 $280k

comments

Want to comment? LOG IN or SIGN UP
TOP 18 Comments
  • LinkedIn swdevl
    I don’t think anyone will hire you directly to google brain without any ML experience. Here is what you can do though. First learn the field well. Stanford has good courses through SCPD. Look at them. They are math hard, but doing ML well needs a lot of maths (stats and linear algebra). If you decide on taking up SCPD, talk to one of the teams in google brain or other ML team and request if they can suggest a small project to be part of the end of course project in one of these courses. If you can pull this off, then you have something to start working with. Do this for 2-3 courses(ML/NLP/CNN), and you now have a good relationship with one or two managers. You have proven yourself at some level, and also skilled up. You chances of switching to one of these teams would increase a lot.
    Aug 9 0
  • LinkedIn swdevl
    However do note that ML requires a lot of math. There are some team that claim they are ML teams, but all they do is call some API of some library that does every thing for them. That’s not ML, that’s just software engineering. And that probably won’t get you anywhere near teams like google brain.
    Aug 9 6
    • New Ogjk
      Can you speak to the salary differences between SWEs who use ML APIs vs actual ML engineers?
      Aug 9
    • LinkedIn swdevl
      Salaries are proportional to the impact you have on the bottom line of the company/org/team, and on the salary policies of the company. Usually in a company, the more impact your work produces (or your management thinks it produces) the more your salary will be. For example , the salary of SRE and production engineers at Facebook would be more than the salary I was earning as engineer in Intel Labs (previous job). I am sure SRE and production engineers at Facebook are smart guys and deserve it, but I don’t think I have the skill set or I want to possess skill set to do that job.
      Aug 9
    • LinkedIn swdevl
      So I guess it boils down to not just salary (I would love getting high salary myself) but also to the kind of work you want to do.
      Aug 9
    • LinkedIn swdevl
      I also think that most of the ML API work ( work where ML means using APIs) are work which won’t produce long term value for you as an engineer. What I mean by that is, a guy with 2 years of experience doing that can easily do the same things as a guy with 10 years of experience doing that.
      Aug 9
    • LinkedIn swdevl
      And as for the original comment above, all I meant was that teams like google brain work on creating ML models etc. Just learning to use existing models or APIs won’t land OP in one of those teams( OP mentioned in his/her post that those are the teams he is targeting )
      Aug 9
  • Google emc2too
    ML may be a bubble. When it pops it may take 15 years to get back to the same level of investment.
    Aug 10 7
    • Bloomberg m5xk27w602
      So why wouldn't they pop again?
      Aug 11
    • Google emc2too
      They could
      Aug 11
    • Bloomberg m5xk27w602
      Basically, all of software is currently in a bubble. I don't understand why this means one should not pursue ML, although I guess it ML jobs could be hard to come by if there were a recession after I completed an MS program
      Aug 11
    • Google emc2too
      ML is bigger bubble because a bigger percentage of the people in ML work on projects that do not have a clear path to monetization.
      Aug 11
    • LinkedIn shillz
      Plus calling ML apis is not ML work... Due to bubble most jobs involving data cleaning or writing python scripts calling ML are sold as ML jobs which actually may be not that great for career in the long run
      Aug 11
  • Google nanithefu
    Curious about this as well. Can any SWE do ML or is it super competitive
    Aug 9 0
  • Google / Eng
    Lπ SWE

    Google Eng

    PRE
    Amazon
    Lπ SWEmore
    🍿
    Aug 9 0