went through the Snowflake frontend interview process last month for a senior FE role on their core product UI team (the Snowsight web app). it's a full product team, not just a dashboard layer on top of APIs, so i was curious how they'd calibrate.
the loop: recruiter screen take-home: build a small data table component with sorting, filtering, and virtual scrolling. 3-day window. technical screen: code review of the take-home + one new live coding problem onsite (4 rounds): two technical, one system design (frontend-specific), one behavioral
take-home: the virtual scrolling piece was the interesting part. they specifically called out performance. i used a windowing approach and made sure to memoize correctly. they also cared about accessibility, which surprised me but shouldn't have.
live coding: the new problem in the technical screen was a debouncing question. they wanted me to implement debounce from scratch, which is fine, but then they extended it: cancel the debounced call if a prop changes. felt like a real Snowsight problem.
frontend system design: this was interesting. i had to design the frontend architecture for a complex filter builder (like the kind you'd use to build a SQL WHERE clause with a UI). they wanted state management strategy, component decomposition, how i'd handle deeply nested filter trees. less about infrastructure, more about product architecture.
behavioral: competency-based. they asked a lot about how i collaborate with design and data. working in a product that sits on top of a data platform means FE engineers have to care about query performance, not just render performance.
overall: harder than your average frontend loop. if you're coming from a pure UI/CSS background without strong data structure knowledge you'll struggle with the system design round. the take-home was a genuine bar-raiser.
offer: senior IC, SF hybrid. base was around 195k. i accepted.