In a world of instant transactions and digital-first services, a security breach can happen in seconds. Legacy fraud detection systems, which rely on batch processing, are no longer enough. This hands-on workshop dives into the world of real-time security analytics. Our time to insight is drastically improved with the use of Event Streaming with Kafka. You will learn how to setup your own Kafka cluster, Kafka producer, and build a powerful fraud detection Kafka consumer using an unsupervised AI model for anomaly detection. By the end of this session, you will better understand how you can use AI and Event Streaming to help you combat fraud at the speed of data.

We will setup a local Kafka cluster using docker, then create from scratch a Java Spring Boot application. The application will use Spring Kafka to produce events to the local cluster. Our application will also function as a consumer for purposes of the workshop. The consumer will use the Smile framework and an Isolation Forest unsupervised learning algorithm to look for anomalies based upon a trained model we will provide. This will allow the detection of anomalies and the appropriate reaction to them.