Facebook Infrastructure Data Scientist phone interview
This interview is with a data scientist over the phone presumably to determine whether to invite me on site or not. What questions should I expect? Can they ask any random data science question or from a usual set of questions related to particular topics (say SQL and/or data munging with Python/Pandas) for the phone interview? If they can ask anything, doesn’t that imply that your chances for on-site interview is pretty much at the mercy of one person and how reasonable their questions are?
Expect 1 coding question, using SQL or Python. Mostly SQL. Know joins, sub queries, aggregate functions. Expect some questions about basic probability sampling/distribution & possibly some simple stats. E.g, how can u tell of two sets of data have the same underlying distribution..Good luck. Luck is key. If you don’t pass, could be because the individual interviewing you gave u a hard time. Could also be because of your incompetence, if you don’t know an area
Thankyou so much for your reply! This is really helpful. The recruiter told that I’ll be getting python and sql question both. What level of leetcoding did you do? Should do dynamic programming too? Also Stats example was really helpful. Do you think they can ask about some strange distributions which are not very common? And can you also suggest the sources from where you studied stats :) Lastly, did you get ML questions ? Do they ask us to derive them? Did you proceed to onsite? Thankyou!!!
Welcome. First of, I didn’t get to on-site. There were no ML questions during the phone interview. I was asked the following: suppose you have the x,y coordinates for all the houses in some neighborhood, and there is a cell tower where the signal reaches 500 yards in every direction, how many houses can be covered. Follow up, if you know the average income per house, how would u find the house with max income in each neighborhood. I figured the first part (you log into coderpad and write your solution, I chose SQL) but messed up on the follow up question, which I believed caused me the chance to on-site. I was asked about the distribution question that I mentioned above. I have heard that the infrastructure team tends to ask slightly harder questions compared with the Analytics team (don’t know about core ML team). The sample FB SQL questions on Leetcode (easy to medium) should be a good starting point. I doubt you’ll be asked SW question about data structures/algorithms. You could expect basic hypothesis testing and related questions. The recruiter should have sent you prep materials/links. Look through them. From all the stuff that I have gathered, know your basic and intermediate SQL well. On site, they may ask questions about simple ML concepts, like bias variance trade off, gradient descent, regularization (type I and II), when to use which method, and so on. But I haven’t made it on-site and cannot be sure. I think aside from technical questions, you have to find out exactly what the role requires and focus on that. Do learn about the role as much as you can. You could ask the recruiter, but the recruiter may not necessarily know. If you have LinkedIn connections or friends that work as DS in FB, you could ask them. The more info you have about the role the better. Good luck
How was your interview? I have an interview coming up