Kafka interviews rarely stop at “What is Kafka?” Interviewers push into consumers, consumer groups, rebalancing, delivery guarantees, and how you operated Kafka in production.
Core topics to rehearse
- Topics, partitions, and keys — ordering guarantees within a partition.
- Producers and acks — durability vs latency trade-offs.
- Consumers and consumer groups — parallelism and partition assignment.
- Offsets and commit strategies — at-least-once vs effectively-once patterns.
- Retention, compaction, and schema evolution — long-lived event streams.
Answer follow-ups like “What about the consumer?”
Interviewers often ask a broad question first, then narrow to a component. When you hear a short follow-up like “Consumer?” or “What about retries?”, answer the specific slice: poll loop, max.poll.interval, idempotent processing, dead-letter topics, and observability on lag and rebalance events.
Bring one production story
Pick a single service you owned: ingress volume, partition count rationale, incident you debugged via consumer lag, or a migration from sync REST to async events. Metrics make answers credible — throughput, p95 latency, lag thresholds, and what you changed.
Using a copilot for Kafka interviews
Load your resume bullets that mention Kafka before the session. When a multi-part question arrives in one utterance, a copilot can help you respond to the latest clause — for example, moving from “What is Kafka?” to “Can you explain Kafka consumers?” without losing the thread.