Interview Questions for Data Scientists

Great data scientists come from such diverse backgrounds that it can be difficult to get a sense of whether someone is up to the job in just a short interview. In addition to the technical questions, I find it useful to have a few questions that draw out the more creative and less discrete elements of a candidate’s personality. Here are a few of my favorite questions.

  1. What was the last thing that you made for fun?

    This is my favorite question by far — I want to work with the kind of people who don’t turn their brains off when they go home. It’s also a great way to learn what gets people excited.

  2. What’s your favorite algorithm? Can you explain it to me?

    I don’t know any data scientists who haven’t fallen in love with an algorithm, and I want to see both that enthusiasm and that the candidate can explain it to a knowledgable audience.

    Update: As Drew pointed out on Twitter, do be aware of hammer syndrome: when someone falls so in love with one algorithm that they try to apply it to everything, even when better choices are available.

  3. Tell me about a data project you’ve done that was successful. How did you add unique value?

    This is a chance for the candidate to walk us through a success and show off a bit. It’s also a great gateway into talking about their process and preferred tools and experience.

  4. Tell me about something that failed. What would you change if you had to do it over again?

    This is a tricky question, and sometimes it takes people a few tries to get to a complete answer. It’s worth asking, though, to see that people have the confidence to talk about something that went awry, and the wisdom to have recognized when something they did was not optimal.

  5. You clearly know a bit about our data and our work. When you look around, what’s the first thing that comes to mind as “why haven’t you done X”?!

    Technical competence is useless without the creativity to know where to focus it. I love when people come in with questions and ideas.

  6. What’s the best interview question anyone has ever asked you?

    I’d like to wish for more wishes, please.

I’m always looking for new and interesting things to add to my list, and I’d love to hear your suggestions.

31 Comments on “Interview Questions for Data Scientists”

  1. What was the hardest part of your thesis work at the graduate school?

  2. Brian says:

    My favorite interview questions aren’t questions I ask the candidate.  Instead, I ask myself:
    1. Is this someone I’ll look forward to seeing day after day?
    2. Is this person interested, turned-on, alive at the thought of doing this work?  
    3. Is this person bringing an interesting/unique slant to the situation, or are they just another mini-me?
    4. Is there anything about this person that I don’t want around?  An annoying laugh, whiny voice or bad body odor can make the office an unpleasant place to be.
    5. Is this person able and willing to cooperate, or are they going to make the office a combat zone?
    If the answers aren’t “Yes, yes, yes, no, and yes”, then the answer to this person is “NO”.

  3. I am a big fan of a hack test: good talkers do not always correlate to good coders?

  4. I also ask a couple of questions about they approach to managing complexity (organization, requirements, inherent technical complexity….) and demonstrated ability to learn from adjacent domains

  5. bwtaylor says:

    1) Tell me about an “aha” moment you had solving a problem.
    2) What do you do when you stare at a problem for a while, try the “obvious” things and get nowhere?
    3) Give me an example of predictions upstanding people make that are junk?
    4) What do you do for fun?
    5) Tell me a few technical ideas that you consider amazing.

  6. Baris says:

    If your twin with exactly the same academic and professional background were here for the interview, tell me why I should hire you instead of her? 

  7. Ole Bahlmann says:

    “Tell me of the last time you had completely unexpected results and what you did then.”
    Gives insight into testing and verification processes and expectation and stakeholder management.

  8. Randy Au says:

    What do you like to do to get people excited about some discovery you made?
    it charts, tables, passionate speeches, beer, or blog posts, we all
    have our preferred toolbox for communicating something that’s exciting
    to us.

  9. I would drop the question on algorithms. I consider myself a data scientist, without being enamoured with algorithms.
    Would add: “What exciting questions would you dream to look at in our data?” “Give me a playful question, and a profitable one” [and recruit the one who gives you a question that is both playful *and* profitable! ;-) ]

  10. Beyond the usual Q&A nonsense, my favorite way to interact with fellow data scientists is at the whiteboard. Good scientists (data or otherwise) can clearly articulate their own ideas, build on each others ideas, and are willing/eager to explore uncharted waters. When we interview at @DsAtweet:twitter we ask our candidates to come in with a “brainstorming topic” in mind, but that they should be prepared for the conversation to be very different. Successful brainstorming sessions generate lots of ideas and don’t get too bogged down in the nitty gritty details (which can always be worked out later).

  11. Taha Yasseri says:

    What is your Dream Data Set? If you could have access, what you would wish to?

  12. Is there “someone” in your data that you like? Usually, you can know certain behaviors of your users, or maybe just where their from. When does data get personal?

  13. Charlesmartin14 says:

    Remarkably unprofessional.  

    • Charlesmartin14 says:

      A good interview would attempt to discover if the candidate can do the job professionally and timely with high quality and a commitment to excellence.

      It is unprofessional (and borderline childish) to ask “what they do for fun” or “what is their favorite X”.   It is even worse to inquire about the details (i.e trade secrets) about what they have done (or should have done) for previous clients and employers.   

      • Hilary Mason says:


        I have a set of technical and more typical questions, which I’m not sharing online, that establish professional competence. These questions go beyond pure technical competency.

        I strongly disagree that you shouldn’t ask people to walk you through their previous work. If the interviewee can’t discuss a “trade secret”, it’s up to them to make that clear, but not my responsibility not to ask.

        These questions are also designed to examine communication ability and creativity, two qualities that are important for a data scientist but generally not verifiable on a technical exam.

        I see that you’ve primarily worked for large organizations, but cultural fit is extremely important in a startup, and within legal bounds I’m obligated to ask what I can to establish it.

        • charlesmartin14 says:

          Well I was going to delete my comment because I made it haste, but I can clarify here instead

          If you want someone to discuss previous work, then you should ask them to prepare something specific in advance, such as a redacted summary of their previous work.

          As for being a good cultural fit, most large, successful firms value diversity and manage it well.

          • Kurt Smith says:

            If someone comes for a data science job interview and is unprepared to talk about their work experience beyond platitudes like “excellence” I’d wonder how serious they are.

        • Angus Urquhart says:

          In my role I have a say in which data scientists we hire. I’d lean towards people who do data scientist outside of work or for ‘fun’. I think it shows that the person has a passion for the topic and for learning in general.

          Things like kaggle, coursera or having their own blog seem like good indicators of where people like doing data science outside of working hours.

    • Asdf says:

      Yes, indeed: it makes one wonder why you left your unhelpful sniping comment at all. 

    • Phillip Burger says:

      Good interviews also try to determine fit of the candidate with the culture of the place doing the hiring. 

      Implementing an algorithm in R for a blog post might be fun. Fun, too, may be contributing to an open source project.  
      Either of these two answers would tell me that the candidate likes to build, that I might have a creative person sitting across the table from me.  Creative people are fun.

  14. I think these are great questions – thanks for sharing Hilary.  

    One thing I like to ask data science candidates is about how to approach specific problems/datasets and the tradeoffs that come with different approaches.  For example, showing them a specific data set (or sometimes just describing the data and how it was generated is sufficient) and asking them the different techniques to analyze it, and what might be the downsides to each – for example using neural networks or SVMs, and what might be some obstacles that would come with each.Another great question to ask is about evaluating results and establishing success criteria.  One of the biggest challenges is determining when something is good enough – which often involves more than just simple precision and recall.

  15. nimrodpriell says:

    Great list! I’ve used your (5) consistently, and let it naturally lead to technical questions (i.e, I follow up with “Great idea! How would you implement that?”). This way, I’m sure that when I’m asking about technical details, I’m probing into somewhere the interviewee claimed to have had knowledge of, and not just poking random technical subjects of my choice which may give me the wrong impression about his abilities.

  16. Aditya says:

    One possible question could be – Which is your favorite ML book and why ? 

  17. Chris Mills says:

    Great list in addition to the technical questions.  One of my favorites is “What’s the difference between a Senior Data Scientist and someone who’s not senior yet?”   I’m hoping to hear comments about mentoring and raising the level of the team.  The worst possible answers, in my opinion anyway, would be focused on things like years of service.

  18. […] time ago, Hilary Mason of did a blog post on the sort of questions she asked when she was recruiting data scientists. There was some interesting stuff there, and since then, other people have done similar things via […]

  19. Sean Hull says:

    I love interview question posts. Everyone can learn from them. Lay audiences who just want to know what it is you do, recruiters & hiring managers who want to find the talent, and candidates who are looking for such a position.

    Great post.

  20. gooderm says:

    What’s the best interview question anyone has ever asked you?

    This one.

  21. Sam says:

    Perhaps a generational
    gap here. I don’t know when in my 30 years of developing different algorithm, I had
    a favorite one. I like different algorithm
    for different needs. This seems like much younger people having favorite
    colors, animal, music, cars, and everything else. Tell be about your good project and/or
    bad project are classical questions many people ask. I don’t use them anymore since it
    mostly shows who is better at interview rather than who can do the job better. In
    my experience those came prepared for cliché questions were not always the best

    I have a set of technical question that I will have my
    senior guys go through them with candidates. They are not theoretical questions or text book questions. They are real situation
    technical scenarios. If they find a candidate suitable, then we invite the candidate to one of our session for chat/
    technical discussion or brainstorming about particular topic with one of the teams in a
    casual setting.

  22. jjjjjim says:

    “I want to work with the kind of people who don’t turn their brains off when they go home.”

    I don’t see how this is relevant at all. The question does not specify domain, and if you ask that question, you should expect any answer that falls within the category “fun”. And if you connect “fun” to “brains always on” and expect strong correlation no matter what, then I don’t see why people would like to work with you. I would not.

    Brains need relaxation and it is absurd to expect people to be constantly involved in brain-stimulating activities … for “fun”.

  23. peterparker says:

    Best way to judge the ability of a candidate in this field is to gauge their basic mathematics aptitude, statistical such as mean,media,mode and probability. The question which comes in mind for probability is to ask the outcome of the head or tails after spinning a coin or three coins. Many people struggle to answer such basic questions. On coding, one cannot judge the coding ability in an interview whose time frame is limited to 40 minutes at the maximum. Though mathematics also cannot determine the ability to code, but showing an aptitude towards mathematics will ensure that candidate is right fit to model the data and possess capability to do data science.