I applied online. The process took 5 weeks. I interviewed at Meta (London, England) in Mar 2024
Interview
The process is very structured and well-documented. Call with an HR -> code pre-screening (45 min leet-code style interview) -> 4 panel interviews (2 leet-code coding interview, ML design, behaviour interview). For ML design preparation the best source is Grokking ML design interview preparation course. Behaviour interview is also a very formalised one, the guide in Cracking the Code Interview is a good one. I haven't reached the bar specifically on the coding interviews, unfortunately. A lot have been argued about the relevance of leet-code style interviews, for ML engineers I found them especially far from the real experience but these are the rules of the game. If you are really motivated to land a job there - practice leetcode until you lose your will to live and a bit more. It has nothing to do with your normal software development experience as an ML engineer but this is a special dance you are required to master if you want to work there.
The process itself was organised in a good way, scheduling, career portal, can't complain about them.
Interview questions [1]
Question 1
For ML design it was about a recommendation system. For coding interviews - pick any leetcode-style medium level problem and implement it extremely fast. Be sure to correctly assess O-complexity for computation and memory
Got connected through a friend who currently works at Meta, which fast-tracked me into the interview process. The first round was a coding challenge where I tackled a task involving user-item interaction logs. The subsequent rounds included a deeper discussion on designing an ML system for video recommendations. Lucky for me, I had spent hours on prachub.com prepping, which helped me nail the coding question on similarity computations. Ultimately, I received an offer but decided to pursue a different opportunity. Overall, it was a solid experience.
Interview questions [2]
Question 1
Design an ML system to rank and recommend short-form videos (Reels) in a user's feed, including how you'd choose features, handle cold-start, and pick an evaluation metric.
Applied online. Received a recruiter screen within two weeks covering background and role fit. Followed by a technical phone screen with coding (LeetCode medium-hard, arrays/graphs). Then a virtual onsite with 4 rounds: 2 coding, 1 ML system design (recommendation/ranking system), and 1 behavioral. Interviewers were professional and gave time to ask questions. Results communicated within a week post-onsite.
First stage was a screen round with behavioral and 2 leetcodes, one medium one hard, 15-17 min each. If selected, loop is 4-6 interviews. 2 desgin, 2 coding, 1 behavioral.
Interview questions [1]
Question 1
idiotic questions that can't be answered in depth in 35 minutes design, like "your solution isn't going to work, how will you handle it?" yea no sht this is a baseline bro, wait 15 seconds and ill talk about the optimal one.