Are SQL problems on leetcode enough for Facebook data science interview?
I have gone through all 42 SQL problems on leetcode and can solve most of the easy and hard ones without hints. Some that are labeled medium are tricky. Is this enough to do well on the SQL portion of the interview? Are the interview problems very similar or possibly the same as on leetcode? The interview consists of SQL, Coding/Scripting, and ML/Prob/Stat. Can anyone provide any examples of Coding/Scripting problems for data science interview. This would be for Infrastructure data Scientist position. Anyone from FB care to chime in? Thank you!
Yes. Some of the SQL leetcode Overkill for FB
Nice name lol
Thanks. What about coding/scripting examples? I’m pretty familiar with basic prob distributions, hypothesis testing/t-tests, and basic combinatorics. Anything beyond these? For ML algorithms, will they ask to write pseudo code on say gradient descent?
You should be able to describe GD but I'd call pseudocode bar raising for that role. It would be a fair question for my team at FB but I'd hire someone even with a handwavy explanation
The first time I interviewed for DS and I dropped out of the process after being asked for a the interview because the coding interview was "reverse a linked list".
Thanks StormyDani. Are you at Amazon or FB?
Neither. I was at FB for three years, went back to grad school, tried Amazon and left really quickly. Teaching college now.
Were you a data scientist at FB? Did u interview people for this role? Are most FB data scientists from top schools with PhDs? I’ve got a PhD in Physics and an MS in Math but definitely not from a top school. Must I shine in the interview to have a chance?
Most SQL problems are based off typical day-to-day queries one needs to make for solving some product analytics problem. If you use SQL enough at work and have a decent intuition behind how to solve product problems, that should be enough.
Actually, the SQL problems I faced in my interview were definitely interesting, and required thinking about the problem more than the SQL syntax. I got the first part right, and sort of ran out of time on the trickier part. I got a little nervous too but after the interview I could see the problem wasn’t that difficult. The thing is that people get nervous during the interview and it’s hard to pause and think when you’re on the phone, because even 10 seconds can feel like a long time. I’m definitely able to handle the job and would love to join FB, but now it’s in the hands of my interviewer to finish his review and submit to the hiring manager. I hope I get called for on site. The interviewer was definitely nice and helpful. I hope he would consider the fact that nervousness did play a role. I’m also kicking myself for not mentioning cross validation when exploring models, but again you get nervous and forget things. Interview was yesterday and so far I’ve not heard back even after emailing the HR person. I’m assuming it’s a no, which’s quite unfortunate. But I’m definitely not giving up, and will practice more SQL. BTW, this was for the Infrastructure Team. I’m not sure if the Analytics Team would count as a different role, or if I get rejected then I must wait another 6+ months? Since you work at FB, any thoughts are appreciated!
Wait for a few more days; interviewers take a few days to submit feedback. The recruiter will get back to you in a few days. I am not sure what the cool-off period is but 6 months sounds OK. Also you can just keep applying to other analytics job description as see what happens. Good luck!
Guess I did not do so well myself in the FB interview . How long does it take to get a rejection? Or if you are moving forward. I wish I could have a do-over exact same feeling like you OP.
Sql for Facebook data science is super easy. Focus on stats and their business metrics
Hi, how do you know this? Did you interview with them? What kinds of technical questions were asked?