Finished the loop for a Senior Data Engineer role on the platform team. Six rounds total, which felt like a lot but was actually pretty well-organized.
Round 1: Recruiter screen, 30 min. Standard background, comp expectations, timeline. No surprises.
Round 2: Technical phone screen with an engineer. 45 min live coding in CoderPad. Two LeetCode-style questions, both medium difficulty. One graph problem, one string manipulation. Nothing exotic.
Round 3: SQL deep dive. This one surprised me. Not just "write a query" stuff, they wanted me to talk through micro-partitions, when to use clustering keys, and how to think about query pruning. I did fine but if you haven't actually used Snowflake internals you should spend a week in the docs first.
Round 4: System design. Design a streaming ingestion pipeline at scale. Felt like a standard data infra design interview, they wanted to see how I thought about backpressure and schema evolution.
Round 5 and 6: Back-to-back behavioral panels. Two separate interviewers, STAR format, very explicit about wanting specifics. One interviewer kept asking for a "more recent example" when my first one was from 3 years ago. Take that as a hint.
Offer came 9 days after the final round. Overall felt fair. The SQL round is the one most people seem underprepared for based on the Glassdoor reviews I read.