In the core of Uber's architecture is a marketplace platform, which is responsible for fulfilling requests for rides, eats, deliveries, and etc. To make our marketplace system efficient and intelligent, we need to extract timely and deep insights from our carefully curated data, and make them available for both people and machines to consume in real time.

This talk will discuss how Uber builds its next generation of stream processing system to support real time analytics as well as complex event processing. In particular, this talk will focus on Uber's stream processing systems that supports many types of real-time analysis and forecasting of geospatial time series, as well as discovery and extracting interesting patterns from data streams. The talk will also discuss how we evolved our streaming system's underlying platform from Apache Samza and Spark Streaming to Apache Flink.