Interview process (Mid Data Engineer — Marketing):

• Take-home (HackerRank): 1-hour Python + SQL (offline).
• Recruiter screen.
• Live coding (Python): standard DS/algos + a bit of SQL.
• System design (data): HackerRank whiteboard.
• Engineering Manager: last step.

My experience:

• Rejected at system design with feedback that felt nitpicky for mid-level.
• Panelists couldn’t say which team I’d join; ~5 min left for my questions.
• After a month, I still didn’t know the actual responsibilities or team setup, although they did say I’d probably end up just writing Spark jobs.
• Feedback centered on tool choices (e.g., Flink vs Spark despite the prompt asking for low latency; Redis vs Cassandra when budget “wasn’t important”; dislike of Kafka Streams). It came across as grading stack alignment rather than the reasoning and trade-offs.
• For a mid role, details like exact stack are usually team decisions, not pass/fail items.
• Communication was rough: interviewers often mumbled or poorly formulated questions, then marked me for not replying fast enough or for asking them to repeat/rephrase, even when I answered correctly once the question was clear.
• Contradictory design feedback: when I added concrete details, they called it “too complex”; when kept simpler, it felt like they’d say “not experienced enough.” Expectations were unclear for a mid role.

Takeaway: Felt like “pick our exact stack or you’re out.” Given the time investment, unclear team/charter, and inconsistent questioning, I wouldn’t go through it again.

Coding challenge: complete a class with some methods and use map reduce later.

System design: online/offline feature store.