I applied through a recruiter.
It started with a recruiter screen and a 45-minute tech screening round with two coding questions on string and array logic. They were more like LeetCode mediums. The onsite was four rounds back-to-back:
The Linux round was definitely the hardest part. It started with a basic troubleshooting scenario about a slow server and went deep into kernel internals, syscalls, and network retransmits. You really need to know your tools like tcpdump and strace. Design was about building a messaging platform at scale, and behavioral was standard as usual. For preparation, I brushed up on Linux fundamentals, LeetCode mediums for coding, and also went through some Reddit posts. I also did a mock on Prepfully with an actual Meta PE. That turned out to be a lifesaver because it helped me tighten up my networking answers, which were a bit weak. The systems round is a total filter, I believe, so you have to know your Linux internals inside out. Overall, it's a fair process but very technical.
The role of VFS in the Linux kernel.
The following metrics were computed from 46 interview experiences for the Meta Production Engineer role in Menlo Park, California.
Meta's interview process for their Production Engineer roles in Menlo Park, California is very selective, failing most engineers who go through it.
Candidates reported having very good feelings for Meta's Production Engineer interview process in Menlo Park, California.