quietquit_quincy · 4 replies
interviewing at Anthropic in about 3 weeks for a senior SWE role (not research, more infra/platform side). would really appreciate if people with recent loops could drop what they saw.
specifically curious about: coding round difficulty and style (LC hard? system-design heavy? something else?) how much does the behavioral actually matter vs being a checkbox any topics that came up more than once
dropping my own data when I'm done. this community has been useful and want to pay it forward.
did a loop 6 weeks ago, SWE infra track. coding was two rounds: one was a real implementation problem (think: build a rate limiter with specific constraints, not abstract), one was more algorithm-y but not pure leetcode. systems design was thorough, they asked follow-up questions that actually tested whether I understood my own design. behavioral was a full round, two interviewers, they clearly scored it.
for infra specifically: know your distributed systems tradeoffs cold. they asked me about consistency models and I got into a pretty long back-and-forth on CAP vs PACELC. they seemed to like people who had actual opinions and could defend them vs people who recited definitions.
that's super useful. opinions vs recitation is exactly the kind of thing that's hard to prep for but good to know going in. thank you.
the behavioral at Anthropic is not a checkbox. I've spoken to people on their side and they calibrate against it. if you're weak on behavioral your technical loop won't save you past a certain point. prep those STAR stories, but also prep to defend why you made the calls you made, not just what happened.