How come Microsoft has such a low hiring bar for Data Scientist position ?

New _xkcd_
Dec 5, 2019 19 Comments

I recently interviewed for a full time Data Scientist role at Microsoft. Surprisingly they didnt ask even a single coding question and most of the interviews were around my resume and basic probability.

A couple of my friends also had very similar experience. Based upon this experience I have started to judge the quality of work and expectations from a full time DS.

TC : New grad from a top CS school


Want to comment? LOG IN or SIGN UP
TOP 19 Comments
  • Microsoft QxrV16
    Not surprised to be honest. I know a quite few people who were PM (civil engineering/ system and informational system major) who doesn't code at all later all became Data Scientists. I think it's mainly because there aren't enough trained data scientists in the current market, and so the company is willing to hire people and train them.

    A couple years ago, Facebook also did the same thing that they will hire people with technical background (CS/Engineering degree) and then train them to be data scientists at Facebook because they just can't find enough people.
    Dec 5, 2019 5
    • Microsoft QxrV16
      Yeah - I think data scientists is still a fairly new career - there isn't a clear set of rules/requirements when hiring people to be data scientists. At Microsoft, I have seen people who has all different kind of backgrounds being hired as data scientists (art major who did a data science online certificate). If you are curious, you can do a search on LinkedIn for data scientist at Microsoft, you will likely see most of them have very different backgrounds. And there's nothing wrong about that, it is a new industry with hot demand and Microsoft is willing to find the potential candidates and train them for the jobs.

      I do notice however that the hiring bar for data scientists now is much higher than the past 3-5 years. Now they mostly hire people who has statistics/mathematics/engineering/data science background.
      Dec 5, 2019
    • Axtria m7kd1
      I have a stats background, my friend has DS background. We prepared together for DS roles, but didn't receive attention. We noticed more number of companies prefer CS grads and train them to be DS, like you mentioned rather than hiring someone with DS skills but no CS skills.
      Dec 5, 2019
  • Axtria m7kd1
    My friend interviewed for DS at Microsoft recently. He was asked about ancova modeling, convex optimization, and leetcode medium. Surprising you weren't asked any of it. I think it's team dependent, no standard process. It's unfair but it is what it is.
    Dec 5, 2019 0
  • Microsoft xIdC37
    There is no standard interview at Microsoft. It’s whatever the team decides is needed.
    Dec 5, 2019 0
  • Microsoft MereDork
    Theres no ms standards. And your resume is more important than anything u can regurgitate on a whiteboard anyways.
    Dec 5, 2019 6
    • New _xkcd_
      OQwM06 I think you should go through the Initial post. My views were for a new grad position. According to MereDork he/she has 15 years of experience. I agree with that much experience on belt white boarding isn’t the best evaluation criteria.

      1. Nobody cares about your school - I disagree with this. For a new grad coming from a good school does help in getting through initial resume screening for both internship/job search. But that isn’t the only criteria, if the student has good VERIFIABLE projects on there resume.
      2. There are enough stats online on how much people fake on there resumes. So coming to your point on proving that one can perform similarly in a non-classroom environment has to be via a comprehensive/quality whitebording interview experience.

      3. An instance from Amazon : For initial SDE filtering, amazon gives out an OA in batches. All of your questions from each batch are online with solutions. A lot of my friends had solutions in hand before they even started attempting the OA.
      My point is that OA and resume both are not sufficient (mostly inadequate) to evaluate a candidate IMHO.
      Dec 18, 2019
    • Amazon OQwM06
      I did read your initial post. You're a new grad, so I don't expect you to have much insight into what interviewers want from you beyond what you've experienced. I don't think we have many points of disagreement but there are a few significant ones.

      1. You keep repeating "a good school". Please give me a list of "good schools" that I should accept people from. Does it include international schools with outstanding candidates? What's the gender and racial make-up of that "good school"? Does it graduate people from across the economic spectrum? Why does sitting in a chair being taught X from a "good school" mean you're better than another candidate who sat in a chair and got taught X from a "not good school"? Do you have any statistics to back up this difference?

      I have given 60 interviews between phone screens and in person interviews. I can count on zero hands the number of times the particular school was relevant in my decision. The only school I even remember from an interview was a candidate from "The Ohio State University" because we were joking about how important the "The" was as part of the ice breaker. *MAYBE* if your recruiter has had a lot of success with a few particular schools, they MIGHT have an unconsious bias to get you past the initial resume screening at a higher rate. Beyond that? If you have an accredited degree, that establishes some minimum level of "sit in classroom and take tests" time and nothing more. WHat matters is your experience, leading us to...

      2. Of course people lie. The resume is just a starting point for the conversation. You said you did x project and achieved y result. Great! What is the metric for Y result? WHy did you choose that metric? What were the other options? What are the pros and cons of each metric? When you did X method what did you have to do with your data? Tell me in detail how that method works? Is that similar to this other method you didn't tell me about? What are the differences?

      3. "Initial filtering". Enough said... like I said, at best you get a foot in the door in resume screening/filtering. Beyond that, nobody cares whether you came from what you think is a good school -- in fact they should NOT care because it would introduce lots of bad biases.

      4. Any coding question that can be answered on a white board in the space of an interview is too trivial to represent real work challenges, or too difficult to solve in the time allotted. It rewards people with the time and access to "coding prep" tools that bear no relation to on the job performance. I'd much rather talk about actual work you've done, why you made choices you made, and what you know about what you did (to screen out liars).
      Dec 18, 2019
  • PayPal / Eng

    PayPal Eng

    Hiring bar is low for all lol
    Dec 5, 2019 0
  • Microsoft / Data MoonBase
    What were the basic probability questions?
    Dec 15, 2019 0
  • Facebook / Data

    Facebook Data

    Microsoft hiring bar varies based on team
    Dec 5, 2019 0
  • Microsoft h.a.l.l.
    Hiring bar is low.
    Dec 5, 2019 0


    Real time salary information from verified employees