A Recipe for a Streaming Data Processing Platform on Apache Flink

I'll talk about the infrastructure we built around Apache Flink clusters, ranging from one to hundreds of Flink jobs, and the API we chose to describe the pipelines.