Importance of PhD -vs- MSc in Data Science Roles

New vxFX54
Sep 20 12 Comments

I have been observing that candidates with PhDs( even the ones with mechanical engineering) get the applied scientist or machine learning engineer roles while the MSc grads are stuck with menial data analyst roles, although all of them share the same "data scientist" name. Does not having a PhD restrict one's shifting from data analyst roles to ml roles ( even though you possess much more work experience than recent PhD grads )?
This behaviour is much more rampant in FAANG companies.
Also, PhD guys get paid more.

Does anyone have any anecdotes of MSc grads getting into FAANG companies in ML roles?

All being said, I have immense respect for PhD grads.
Just a tad bit..... overrated.

comments

Want to comment? LOG IN or SIGN UP
TOP 12 Comments
  • Google
    kielbasa

    Google

    PRE
    Amazon, Yahoo, Microsoft
    kielbasamore
    I’m sure the phds think they should be paid more due to the huge opportunity cost they incurred. PhDs will likely make less over a lifetime due to the opportunity costs they incur early on their career. They also had to do a lot of menial work while working in the PhD.
    Sep 20 8
    • Google
      kielbasa

      Google

      PRE
      Amazon, Yahoo, Microsoft
      kielbasamore
      That’s true about the out of pocket cost for many master’s degrees. Even so, 4-5 years of a PhD means not having a tech salary while you’re “off the market.” The PhD stipend is simply a means to not go into debt, you don’t really save any money. Master’s students can start to recoup their losses sooner, though they likely have debt. From a financial perspective perhaps both master’s and PhD are bad ideas.
      Sep 20
    • New vxFX54
      OP
      @fsck @kielbasa Agreed. Though, this a tangent to the post.
      I was more interested whether in knowing whether one should be ok with the MSc and the head start or do a PhD, if the long term career goal is to be in core or applied data science.
      Does the additional PhD outweigh the work experience and headstart you
      get from an MSc?

      Losing $500K dollars for the first five years seems alright if a PhD gaurantees , from what I heard starting at 250K TC and unbounded growth, while the MSc guys are stuck at a 300K TC doing menial data analyst/BI jobs

      Keeping in mind, financial costs are not a constraint.
      Sep 20
    • Google
      kielbasa

      Google

      PRE
      Amazon, Yahoo, Microsoft
      kielbasamore
      Doing a PhD doesn’t guarantee anything. If you already have an MS and are working in the field you are on the right track. All you should need to do is build your data science/ML skills outside of work. You just need to pass the interviews for the job you want. You don’t need a PhD for that. I know someone with math MS who became a data science director by studying data science on the side and then applying for data science positions. The questions you get during interviews don’t require a PhD.
      Sep 20
    • New vxFX54
      OP
      Extremely true and that's the way it is supposed to be. But it's getting extremely hard to even get the initial break into the industry without a PhD. An isolated instance of someone's success would only make you a victim of survivorship bias.
      Sep 20
    • Groupon nfof26
      Definitely not. Get an MSc and multiple ML publications and it is very manageable.
      Sep 20
  • Amazon / Eng Beef Gates
    Core DS, like building production models facing customers, most ppl seem to have phds, especially managers and above, but not without an exception, seen some msc guys up there. Otherwise, most of them seem to be in BI orgs.
    Sep 20 2
    • New vxFX54
      OP
      exactly.
      MSc peeps, I believe, are extremely undervalued for BI roles
      Sep 20
    • Amazon / Eng Beef Gates
      Im one of them msc in bi org, been with multiple companies and thought hard about switching to Product management, which pays much more and more valued, but one thing i learned is that some companies started to figure out how to really use BI/decision support to the extreme, and started valuing them much more than they used to. Theres a huge difference in output and thought partnerships between menial data analyst and an experienced BI professionals with proper analytics/stats background, and they know that these phds will not resort to it and having hard time finding the right talents who fit those profiles. So yeah I’m still on in this ride, and I hope I didnt make a mistake by not switching to the PM world.
      Sep 20