From JAR to DAG: Running Java Tasks Natively with the Airflow Java-SDK

Airflow is Python-first — but production business logic is often Java-first. The Airflow Java SDK bridges that gap by letting you mix Java and Python tasks within the same DAG, without shell wrappers or separate services.

In this talk, we’ll walk through the full lifecycle of a Java task: how the Java SDK is set up, how tasks are defined and packaged as a JAR, how Airflow picks them up and runs them on any executor, and how results flow back into your DAG. We’ll also cover how core Airflow primitives — Variables, Connections, XCom, and logging — work natively in the Java SDK, enabling true cross-language, bidirectional communication within a single pipeline. You’ll see it all running end-to-end in a live demo alongside Python tasks.

If you work with Java services and have ever wished you could orchestrate them without translating everything into Python first, this talk is for you.