I applied through an employee referral. The process took 3 months. I interviewed at Meta (Mountain View, CA) in Jan 2014
Interview
My resume was entered into the system by a friend who currently works there, but after a week and no contact, I applied online to a Data Scientist position.
Data scientists at Facebook have a totally separate hiring process from software engineers.
You have an initial phone screen by a data scientist which will focus on your 'analytical ability.' For those of you who (like me) have no idea what that means, it means a tiny bit of coding/scripting in a language of your choice while you work a reasonable, made-up data science scenario. They'll give you pretend access to a pretend database of information, have you write a few queries, give you fake data for your output, and have you debug plausible scenarios for that fake data.
I received word rather quickly (two days later) that I passed the phone screen and would be invited to Mountain View for a day of interviews. I scheduled those interviews for 3 weeks down the line.
Interviews at Mountain View are grueling, not because of their technical difficulty, but rather because of the interview setup. I was interviewed in a tiny closed cubicle no more than 8 feet x 8 feet; room for two one-seater couches and a tiny table. The wall was a whiteboard. There were 5 back to back 30 minute interviews, and while the interviewers were apparently supposed to ask if I needed water or a bathroom break, they often forgot to do so. The next interviewer was waiting right outside when the last interview ended. After we covered all of the technical content (about which I signed an NDA, so I unfortunately will not share the details of that), I had about 120 seconds to quiz my interviewer about what data science is like at Facebook.
I may have earned brownie points with one on-site interviewer for stopping him when he started asking me the same question that I had had during my phone screen. He thanked me and changed to a new problem.
I studied for the Data Scientist interviews by:
a) coding in python (which I do for my job; they were happy to let me code in python for the on-site interviews)
b) reviewing Stanford's online statistics 101 class
c) doing a few 'hat trick' type probability puzzles
I was well prepared for their interview questions.
I heard back from my recruiter 1 week after on-site interviews and received a generous offer with a fungible 2-week acceptance deadline.
The Interview Process is very structured -
First Tech Screening round - 45 mins (usually can extend a bit depending on the interviewer)
- 2 SQL Questions ( Medium to Hard ) - based on Joins
Full Loop - 4 rounds 45 mins each.
- SQL
- Behavioral
- Analytical Execution - stats & prob, A/B testing, case study
- Analytical Reasoning - Case study
Interview questions [1]
Question 1
Questions on Bayes Theorem, Probability distribution, etc.
A friend who worked at Meta recommended I apply, which led to a direct connection with the hiring team. My interview journey included a technical screen focused on product metrics and a behavioral interview. The final round faced me with a practical question about measuring success for a new push notification feature. Funny enough, I had found similar questions in the company-specific prompt section on PracHub, which helped me approach it confidently. Ultimately, I received an offer but decided to decline it for personal reasons.
Interview questions [1]
Question 1
How would you measure the success of a new push notification feature on Instagram? Walk through your primary metric, guardrail metrics, how you'd design the A/B test (randomization unit, sample size, duration), and the tradeoffs between short-term engagement lift and long-term retention impact.
I applied online. The process took 6 months. I interviewed at Meta
Interview
Completed 3 rounds of the process, which includes the initial recruiter screen, technical, full loop, and team matching.
Couldn't move past the full loop interview. The interview was very engaging, and I actually enjoyed working through the cases. No crazy questions.