I applied through a recruiter. I interviewed at Meta (Menlo Park, CA) in Jan 2026
Interview
got a dm from a meta recruiter on linkedin. first was the recruiter screen - pretty chill, just walking through my background and ds experience. then a technical screen which was split between sql and a product case study (identifying metrics for a new feature).
the onsite was 3 main rounds and definitely the most intense part. first was the analytical reasoning round - this was pure product case work, like how would you measure the success of meta verified or what to do if a specific metric tanks. then the statistical execution round which was much more technical - brushed up on a/b testing, probability, and how to actually calculate p-values or sample sizes. last was leadership & drive, which was behavioral stuff about handling disagreements with eng or how you drive impact without authority. honestly i was struggling with the stats part and structuring my case answers earlier but i did some mocks on prepfully with a meta ds coach. it helped a ton with making my logic sound more 'meta' and not just rambling lol. also skimmed through reddit a bit. got the offer shortly after. definitely practice sql under time pressure and get your product frameworks down!
Interview questions [1]
Question 1
What are some potential biases that could occur in a study measuring the effectiveness of a new feature, and how would you account for them?
I applied through a recruiter. The process took 6 weeks. I interviewed at Meta (Londra, Inghilterra) in Apr 2025
Interview
Full loop process for Product Data Scientist. 4 core interviews:
- Analytical Execution: stats / probability (eg probability of independent events). I was also asked to calculate a z-score by hand.
- Analytical Reasoning: case study on a product question. Mine was on website tagging for ads.
- Technical Skills: Live SQL coding.
- Behavioural: standard "tell me about a time" questions
Interview questions [1]
Question 1
The one thing I wasn't really prepared for was calculating the z-score by hand. Other probability questions I did fine on, but I just told the interviewer that I wasn't going to be able to remember the formula for calculating the z-score.
I applied through an employee referral. The process took 2 months. I interviewed at Meta
Interview
- Recruiter screen
- Coding + product question screen
- 4 rounds of 1 SQL, 2 product case studies, and 1 behavioral
Overall very structured and plenty of online resources to help prep. Did take a significant amount of time though, allot a few days off work for interviews at least
Interview questions [1]
Question 1
Let's say metric X goes down, how would you investigate it?
[I would look into A, B, C]
followup: Okay let's say A and B look fine, but C looks off. How would you address it?