Problem
You are designing a streaming pipeline where some events arrive after their event time. Those late records can invalidate previously computed windows, aggregates, or fact tables. The goal is to keep downstream data correct while limiting unnecessary recomputation.
What this tests
- Event-time processing and watermark design
- Idempotent writes and deduplication
- Repair strategy for very late records
- Data quality controls for corrected outputs
Key design signals
Common trade-off·Longer watermark vs lower latencyRecovery pattern·Targeted backfill instead of full replayPrimary challenge·Correctness under out-of-order arrival
Practicing as: Data Engineer interview at TuroHi, I'll play your Turo interviewer for the Data Engineer role. Candidates describe these interviews as often stressful and moderately difficult, so expect me to be direct and to the point. Take your time with the question above and answer like we're in the room.
You are practicing as a guest. Sign up free to get your answer graded with AI feedback. Your draft stays right here.
