Real-Time Data & Messaging at Scale with Apache Flink & Apache Kafka

This meetup showed what a modern real-time data architecture actually looks like - from the compute layer to production systems. We explored the role of Apache Flink as a core compute layer, enabling data to be processed at ingest for better latency, lower costs, and higher flexibility. Through Wix’s real-time feature store, we saw how combining Apache Kafka, Apache Spark, and Aerospike allows handling billions of daily events with near real-time updates. We also learned how rethinking messaging infrastructure - moving to a gRPC-based proxy architecture - can significantly reduce costs, improve scalability, and simplify developer workflows. The takeaway: modern data systems are real-time by design, compute early in the pipeline, and are built to optimize both performance and developer experience.