Did a data scientist loop at Linear in early 2026. They're a small team and don't hire a lot of DS roles, so the process is a bit less standardized than what you'd see at a larger analytics org. Here's what I saw.
First, a note on how they use data. Linear's product is a project management tool used by engineering teams. Their internal data questions are mostly around product analytics: how features get adopted, where users churn, how different workflow patterns correlate with retention. It's not an ML-heavy environment. If you're coming from a modeling-heavy role, be ready to reframe yourself around decision-support and product analytics.
SQL round. This was a live coding exercise with a realistic schema: something resembling issues, projects, teams, and activity tables. Questions included: A window function problem (find the first activity event per user per project) A retention cohort query (week-over-week active users) A self-join (find pairs of issues that were linked within 24 hours of each other)
Difficulty: medium. Not the "write a 4-level CTE" stuff you get at data-heavy companies. But they wanted clean, readable SQL, not just correct output.
Case / product sense. Got a product scenario: engagement with a specific feature was declining despite new users increasing. Walk through how you'd diagnose it. This is a pretty standard product analytics case but they cared about the structure of my reasoning, not just the SQL I'd write to investigate.
Stats. Brief. One question about experiment design: how would you run an A/B test on a feature that only affects a small segment of users. I talked through power analysis, minimum detectable effect, and the option of a holdout design. They seemed satisfied with the framework.
What they didn't ask. No ML modeling questions. No Python. No pipeline design. If you want an ML research environment, this is probably not it.
Overall. For a DS candidate, the signal they're optimizing for seems to be: can you help a small team make better product decisions, fast. Clean SQL, clear thinking, product intuition.