I had a video interview with a Data Scientist. The interview lasted about 45 minutes and consisted of one technical coding (SQL) question and a product question. The team member gave me time to ask questions at the end.
Interview questions [1]
Question 1
Given a database of posts and a database of comments on those posts, how do you determine how many conversations are happening in the comments per post on average.
I applied through a recruiter. The process took 1 day. I interviewed at Meta (Seattle, WA) in Apr 2020
Interview
FB Recruiter reached out to me via Linkedin and we set up a meeting. It was a phone screen by Recruiter for 30 minutes, she was nice and helped me a lot to understand the details of the projects and scope of the work
Interview questions [1]
Question 1
She asked basic statistics,probabilty questions such as bayes' theorem and hypothesis testing
I applied through an employee referral. The process took 6 weeks. I interviewed at Meta (Menlo Park, CA) in Apr 2020
Interview
The process was extremely well-managed, and the recruiters were very prompt and helpful. They do a great job of preparing you for each step of the process.
The first round is a telephone screening which was around 10-15 mins long. I was asked a very open-ended question about how I would broadly analyze the drop of key performance metrics and 5 simple questions on SQL.
The next round, which was the technical screening, took place around 3 weeks later. A similar broad and open-ended product question was posed, and I could not help but feel that the only thing that was being evaluated was my ability to think in a structured manner. This was followed by a couple of relatively simple queries where I was asked to write a code to get a frequency distribution and a simple case-when application.
The onsite round took place online. The first 2 rounds were very similar to the first round in the technical screening in terms of the open-ended nature of the questions, and the evaluation criteria still seemed to be more of how you walked the interviewer through your thoughts without getting overwhelmed or sidetracked. The technical round was SQL again (though there was no restriction on the language I wanted to use), and the problems where definitely slightly more challenging. You need to be quick on your feet with SQL, I was asked what sounded like a very simple question which actually involved a cross join and some creativity on the spot. The last round was a quant round that revolved around Bayes' theorem, Power of tests, Type 1 error/type-2 error, and imbalanced classification. Intuitive explanations of the math were just as important as being able to solve the problems.
All the interviewers were extremely pleasant and conversations. Nobody seemed like they were 'out to get you', but were there trying to actually assess whether you could think coherently and whether they would be willing to work with you. It would have been a great experience even if I had not received an offer, but they were kind enough to extend one to me within 3 days of the onsite round.
Interview questions [1]
Question 1
1) Provided a table with user_id and dates they visited platform, find the top 100 users with the longest continuous streak of visiting the platform as of yesterday.
2) Provided a table with page_id, event timestamp and a flag for a state (which is on/off), find the number of pages that are currently on.