Exactly-once transaction ETL
Design an exactly-once streaming ETL that consumes ~120k events/sec of financial transactions from Kafka, enriches each with reference data and FX rates, and writes aggregated per-merchant hourly revenue into both an OLAP warehouse (for analytics) and a low-latency KV store (for a live merchant dashboard). A consumer crash, a rebalance, or a redeploy must never double-count or drop revenue, and the dashboard must converge to the same totals the warehouse reports. Reads are heavy on the KV store (10k QPS); the warehouse is append-mostly.
What a strong answer looks like
Clarify scale and constraints first. Propose a clean component breakdown, then go deep on the hard parts (data model, bottlenecks, consistency, failure modes) and name the trade-offs you are making.
Clarify5:00 left
Estimate5:00 planned
Design15:00 planned
Deep dive12:00 planned
Failure8:00 planned
Which questions mattered is sealed until you submit. Telling you now would just be handing over the edge cases.
Run or narrate your approach, then ask the coach.