Meta Data Scientist Interview Experience — SQL, Fake Accounts, and an AI Ad-Creation Case Study in the Onsite

Meta·Data Scientist·Feb 2026
Onsitemedium

I just finished my virtual onsite a little while back. I ran into a few questions that other people have already posted about, but with slightly different follow-ups, so I wanted to share.

  • SQL: active_ads. The interviewer was very organized and gave clear instructions. My second question was to find people whose total spend last year was over 1k, then look at what share of their spend this year that group accounts for. The last question was open-ended, centered on whether growth in one creation source was actually caused by a decline in another source — the SQL for it ended up being pretty random, though. Based on where I took the discussion, the interviewer went along with it and had me write a query to exclude certain categories of advertisers. So my advice is to focus on knowing the tables well rather than fixating on the exact questions other people report getting.

  • AE: fake accounts. This one was fairly standard. It ended up being the round I was most nervous in, and I blanked for a bit in the middle — hoping that doesn't sink me.

  • AR: not a plain ads-ranking-algorithm question. Instead it was about an AI-assisted ad-creation feature they want to roll out to advertisers — how would you evaluate it, what metrics would you use, how would you set up the test, and so on. The trick is not to fixate only on the creation piece itself; you also need to think about the broader impact. So it counts as a half-new question.

  • BH: fairly standard behavioral questions.

Published

Curated and edited by PracHub

Practice the questions from this interview

Discussion

Sign in to join the discussion. The author is notified of every comment.

Loading comments…

Interview at a glance

Company
Meta
Role
Data Scientist
Rounds
Onsite
Difficulty
medium
Interview date
Feb 2026
Questions from this interview
2 questions

Real Meta interview experiences

First-hand reports from Meta candidates — the rounds, the questions they were asked, and how it went.

All 125 Meta interview experiences