Scale AI · Primly Community

Collecting recent Scale AI interview loops. Share what you've seen.

market_realist · 5 replies

interviewing there in a few weeks for a senior eng role and the public data is a little thin. want to get a clearer picture of what the 2025-2026 loop actually looks like, especially whether the system design round has evolved with all the AI tooling they've added.

if you've been through the loop in the last 6 months, drop what you can here: role, level, how many rounds, what types, roughly how long it took. no need for super specific details, just trying to build a picture.

also curious if anyone knows: do they do live coding or async take-homes for the first screen? i've seen both mentioned.

5 replies

corp_refugee

went through their loop about 4 months ago for a staff eng role. five rounds total: recruiter screen, async coding challenge (timed, 90 min, on their own platform), then 3 live rounds in a day. no separate system design, they weave it into the later rounds more naturally than most. coding challenge was python-heavy, unsurprisingly. felt like big tech process but faster.

infra_ines

just did their infra eng screen two weeks ago. first screen was live coding, not async. interviewer was senior eng, asked me to implement a basic rate limiter and then extend it. classic. didn't get to the full loop yet but they said 3 rounds after this. waiting to hear.

market_realist

interesting that you got live vs. async, sounds like it varies by team or role. how was the interviewer? scale interviewers have gotten mixed reviews on being present vs. zoned out.

infra_ines

genuinely engaged, asked good follow-up questions, seemed to care whether i was thinking out loud. better than most.

staff_steph

for staff level they added a systems thinking round that's less about whiteboarding and more about 'you're the TL for this ambiguous thing, walk us through how you'd scope it.' not a standard design round. closer to an exec brief. that one threw me a bit on first pass.