I applied through an employee referral. I interviewed at Meta (Menlo Park, CA) in Jul 2017
Interview
Had a phone screen with a recruiter that led to a video interview. Video interview covered building models to solve specific FB problems (are people connected/not), some ML, feature engineering, identifying pitfalls in a model and data leakage, also covered some SQL (though I was given the option to use Python as well). Onsite interview was a series of five interviews broken up by lunch with a DS, and a brief meet and greet in the morning. The topics where mostly focused on product analysis and A/B testing (if we have a product and see some trend in different users make a hypothesis about why we see that trend and test it). One interview was probability and data focused with a couple college level probability questions. One interview was focused on data manipulation in the language of my choice. And the last interview was SWE with a CS type algorithm development problem.
Overall, I would say that a great deal of the interviews did not focus on what I (as a working Data Scientist) do on a daily basis, which is primarily building ML models. There was a definite bias towards A/B testing and almost all of the interviewers brought in situational product based questions.
During lunch the DS I was with was very open and forthcoming about what working at Facebook is like and the types of problems they tackle on the job. He was also very frank about the lack of ML work done by DS's at Facebook and basically said they have one of the best ML engineers in the world working there and he built a suite of tools that are better than anything you will build, ever. As far as his interaction with ML went, he just collected the data he needed and passed it into one of these tools.
In the end I did not receive an offer, though I was provided some vague feedback. My technical skills where very strong, but product awareness and analysis ended up hurting me. Full disclosure I do not have/use Facebook or Instagram and though I disclosed this before beginning the process I felt that it had an impact on my interviews.
Interview questions [2]
Question 1
We have a product that is getting used differently by two different groups. What is your hypothesis about why and how would you go about testing it?
I applied through a recruiter. I interviewed at Meta in Apr 2017
Interview
The recruiter was super friendly, well-organized and always communicate in a timely manner.
First round was a 30min phone call with the recruiter.
Second round was a one hour tech video interview.
Third round was an onsite.
I applied through a recruiter. The process took 3 weeks. I interviewed at Meta (Menlo Park, CA)
Interview
Note: the role was for the Analytics role
Surprisingly very simple questions onsite. SQL, probability, etc. Questions seemed too easy for a real data science role. And the interviewer was kind of a jerk.