TikTok’s real-time recommendations engine, Monolith, is so good it has been described as "digital crack".  In this webinar, we will show you how to build and operate a real-time feature streaming pipeline that writes at 1m ops/sec - TikTok Scale - and concurrently reads data at 100s of thousands of ops/sec.

Like TikTok, we will use Apache Flink as the stream processing engine, and Hopsworks as the real-time Feature Store that will store data used for real-time recommendations. We will look at how to make the whole architecture highly available and how to elastically scale up/down this architecture.