Till Rohrmann

Build AI agents as simple code—keep your favorite agent libraries, model human-in-the-loop steps, handle interruptions cleanly, and get deep, step-level observability out of the box._x000D_ _x000D_ We’ll show how a log-as-runtime approach turns tool calls and LLM steps into resumable actions, so retries, crashes, and long waits never lose progress or double side effects._x000D_ _x000D_ You keep the ergonomics of RPC-style development while inheriting Kafka-style durability and scalability for long-running agent loops._x000D_ Because the runtime is serverless-friendly, these agents scale to zero, fan out massively, and resume instantly—without custom worker pools or glue code._x000D_ _x000D_ You’ll leave with a compact blueprint to ship production-grade agents and agentic workflows quickly, with resilience and operability built in.

With every release Flink is getting closer to become a truly unified batch and stream processor. This unification is not only a big effort for the API layer but it poses also great challenges for its runtime.

Processing batch and streaming data with a single, unified engine have been the goal of the Apache Flink community for several years. Over the last months, the community has made good progress towards realizing this vision.