How do you compute real-time features (streaming aggregates)?

hard

Answer

$(1) Streaming engine (Flink / Spark Streaming / Materialize) consumes events (Kafka).$
$(2) Windowed aggregation: tumbling / sliding / session windows.$
$(3) Write result to online store (Redis / DynamoDB).$
$(4) Handle late-arriving events with watermarks + allowed lateness.$
$(5) Exactly-once semantics via checkpointing.$
$\mathrm{purchases}_{\mathrm{last}}\mathrm{hour}$
$Latency: sub-second.$
$Complexity: state management + backfill from historical events for training.$

Check yourself — multiple choice

Random
Streaming engine (Flink/Spark/Materialize) + Kafka events + windowed agg + write to online KV + watermarks for late data + exactly-once checkpoints; sub-second latency + backfill for training
Just batch
Not real

Real-time features: streaming + windows + online KV + watermarks + exactly-once.

#features#infrastructure

215 interview questions in this topic.