1. Recruiter call - they prep you really well on what to expect through the whole process. 2. First technical round - SQL and business case together - do the Meta problems on stratscratch and you should be fine. Quiz yourself on different products/features at Meta and how to improve them 3. If you pass above, another recruiter will talk you through how to prep for the onsite. Their resources are very comprehensive; Nothing in the interview should really be a surprise. There's also a monthly prep session led by current data scientists to give candidates a live session of how to evaluate a business problem. Despite being a large company, I felt like Meta took a lot of care to make the interview process very personable and meaningful beyond the interview itself. Prepping for it definitely sharpened me for other interviews and helped train me to analyze problems more effectively. I'm very fond of the work they've put in and the process they have.

It kicked off with an initial phone screen that jumped straight into technical questions, so you really have to be warmed up—especially on A/B testing: framing the hypothesis, picking primary metrics and guardrails, walking through power/MDE, exposure and unit of randomization.