To CS PhD or not?

VMware
kanta

Go to company page VMware

kanta
Feb 6, 2021 20 Comments

I’m turning 26. Thinking of a PhD in CS in a non-ML field (systems/networks/security). Assuming I can get into a top 10 school (not guaranteed, but I have a masters from a top 4 school in this field - may help). Primary drivers:

1. Personal growth: not working on small features, but big problems. Research is more fun than industry in general. These may be the sharpest years of my life - so better to do it now. Although I don’t want to go into academia.
2. Career growth: Increasingly I see very top level positions in technical leadership to have PhDs. I understand it’s not a requirement.
3. Entrepreneurship - obviously a PhD is not needed for this, but I see this increasingly as a trend in this domain. I think I’m more likely to engage in this in a PhD program than working on a visa in bay area at a large company 🙃
4. Immigration: PhD doesn’t guarantee immigration, but it can help, especially if I do it from a good place and do well. (because of my nationality, I have no realistic chance of getting a green card in my life time) — this is a minor reason, but it is there.

Looking for some advice or constructive feedback from people who did and didn’t go down this path

Potential cons I see are:
1. Lost tc short term
2. Mental health (PhD can be frustrating, uncertain)
3. Possible no impact on career - I already work this folks who have PhDs lol
4. Age: If start when I’m 27, I’ll be 32-33 when I finish.

comments

Want to comment? LOG IN or SIGN UP
TOP 20 Comments
  • Google
    test@

    Go to company page Google

    test@
    You have chosen the worst possible reasons to do a PhD. For immigration, a PhD won't "help" you. You need to have 200+ citations, best paper awards (yes, the bar is rising - I had an EB1 RFE with both of these things) and be known as a world class expert. The paper publishing rut will crush you and you have no idea how hard it is. Having a PhD will not make you attend more conferences. Also, only ML rules industry research. There is no "research" for non-ML fields. Even if there is, your growth will still be based on impact. Google doesn't give a damn if you can publish papers. Google is not a university, it is a business. You need impact and that has nothing to do with a PhD. To give you another data point, all PhD new grads join Facebook E4 with the title "Research Scientist" and non PhD grads are just SWE title but they do identical work. To be a part of FAIR or Google Research (to do actual research and not just be a swe), you need to be an ML person and have a profile good enough to be a TT professor.

    If you don't want to go into academia (not even have an ambition to), PhD is the worst thing to do. You really need to introspect hard why you want a PhD.
    Feb 6, 2021 6
    • VMware
      kanta

      Go to company page VMware

      kanta
      OP
      Yes, paper publishing pressure is real. I could see this among PhD students during my masters. Though systems work tends to be lower velocity than ML in general.

      @asshole! Given how things turned out, would you do it again?

      As I said my biggest reason is how one have more freedom to pursue rather large and creative problems, rather than work on smallish enhancements to an existing code base at a large co.
      Feb 6, 2021
    • Google
      asshole!

      Go to company page Google

      asshole!
      Interesting question Kanta. For me yes.

      But that’s also because I started with a lot of personal issues that required a lot of independent thinking and free time to make sense of. If I had directly worked in a corporate, I would grow old very quickly completely starved of any sense of clarity about myself and my capabilities that comes from deep introspection and free thinking.

      But now I feel bad because many people who didn’t do PhD are 1-2 levels higher than me at the same or even younger age.

      So it’s a mixed bag. But overall, answer is yes. As Steve Jobs said you can’t connect the dots going forward, you have to let your intuition be the guiding force. I hope there will be a point in future where I will see very clear material benefits as well that will come from all the meta learnings that I had in PhD.

      But then again Steve Jobs never finished a degree.. so whatever 🤷‍♂️
      Feb 6, 2021
  • Upstart
    lkkUkj

    Go to company page Upstart

    lkkUkj
    I have a PhD. You do a PhD because you want to do a PhD. You don't do a PhD for career growth. The only exception is if you need an advanced degree to immigrate.
    Feb 6, 2021 5
    • Google
      asshole!

      Go to company page Google

      asshole!
      Yes true. Google is fking awesome in research and I am indeed an asshole.

      But oracle here is a sore loser with nothing going on in his life. He is one of those people with such IQ levels that believe that Eiffel Tower doesn’t exist because they have never seen it lol 😂
      Feb 6, 2021
    • VMware
      kanta

      Go to company page VMware

      kanta
      OP
      umm, @oracle, get out of your bubble. That’s not true at all.
      Feb 6, 2021
  • Google
    asshole!

    Go to company page Google

    asshole!
    I also went for a PhD later in life ( around your current age) and then did really well in the PhD. Afterwards joined Google as SWE.

    Here is my take on it:

    1. Money/yoe/promo etc - you will lose a lot of time if you come back to SWE. If you go to research such as academia or research scientist positions, you will still lose time, but it wouldn’t feel that way because you couldn’t get there anyway without a PhD.

    But if you become a SWE after that, you will definitely feel the burn. Especially if you compare against people who are doing well, as they will be most likely younger than you and 1-2 levels higher. In established companies, no matter What you do you can’t catch up very quickly, and comparatively may never be able to catch up as some of them will keep growing as well.

    2. Unless you go into a clearly marked research field/position, the only benefits of PhD are basically your meta level learnings. You will learn traits like resilience, independence, starting on a problem from first principles, critical thinking, creativity, ingenuity, technical writing, presenting and selling ideas.

    All these are assumptions based on a good PhD ( my PhD won university level award ).

    Now the catch is that you will see many people in industry have learnt the same but in much less time, like 1-2 years and now they are ahead of you 🤷‍♂️

    But then some people got stuck in the industry and didn’t learn anything.

    So in short, PhD from a good school is very likely to teach you these very useful skills. But when you come out and join the industry, you will be behind many people who got the same skills in industry while also making money and getting promoted. But you will be ahead of people who got stuck or never developed these skills.

    It’s upto you if you see glass half full or glass half empty.
    Feb 6, 2021 0
  • Google
    Moonboy_

    Go to company page Google

    Moonboy_
    The lost TC is more than just short term. 6 years of lost TC and promos and YOE is huge.
    Feb 6, 2021 0
  • Uber / Eng
    lol0

    Go to company page Uber Eng

    lol0
    Just do PhD fellow Chinese/Indian and get GC.
    Feb 6, 2021 0