Meta Data Science interview questions

Recruiter reached out to scout for interest - he was extremely helpful, informative, and encouraged me to take my time to prepare with thorough material. FB wants you to be as prepared as possible with no surprises/tricks to the interview to pick out those who are not only the most qualified but also the best prepared. Scheduled the initial interview around ~2.5 months out. Video Interview - 45 minutes with a Data Scientist. He opened with a quick rundown of my background and why I want to work for Facebook, but the bulk of the interview was technical - the first half being SQL on Coderpad and the second being product interpretation. I passed and received confirmation that I would advance to the onsite on the same day. On Site Interview - 4 30-minute rounds testing various topics surrounding applied data, product sense, probability & statistics, and SQL (or Python/R), with a 45 minute lunch in between with a data scientist who is there to help you answer any questions (not part of the interview. Zero questions on behavioral or anything regarding my background. I was not extended an offer and was communicated the rejection shortly after. Overall, a very pleasant experience - FB is a top notch company with best practices, though of course, individual experience can and will widely vary. For on-site, it's very important that you map and diagram your thoughts on the white board. Most questions appear as though there is a specific answer in mind that they are looking for, though the way you get there might be flexible. The best way to prepare is to think through the different scenarios and products and figure out a way to build a business case around measuring some metric or improving the product. Expect lots of pushback with detailed, probing questions during your answers.