Data engineer working on streaming data modeling: how to model events, state, and time in Apache Flink and Kafka so results stay correct under late data, replays, and failures.
Now building
- stream-patterns: a catalog of streaming data modeling patterns. Each pattern will ship with a reference model, a Flink SQL implementation, and tests. (in progress)
- streamproof: a correctness test harness for Flink jobs, using seeded chaotic input, fault injection, and an oracle comparison. (in progress)
- helion: streaming operators with proven error bounds (Count-Min, Space-Saving, pane windows) for Flink and Kafka Streams.
Upstream
- Apache SeaTunnel: #11817, declarative option validation for the Email connector (merged).
Also
- metabase-jalali and dimension_date: Persian (Jalali) calendar support for BI tools.
Stack: Apache Flink · Kafka · Iceberg · SQL · Python

