ksqlDB is an event streaming database that enables real-time data processing against Apache Kafka, providing SQL-like semantics for reading, writing, and processing streaming data in real-time at scale. To extend its functionality, ksqlDB supports creating user-defined scalar functions (UDFs) and user-defined aggregate functions (UDAFs), which can be built using Java classes with additional dependencies if needed. These custom functions allow users to add new capabilities to their ksqlDB statements without modifying the existing syntax. Building a UDF is straightforward, requiring only one Java class, minimal wrapper code, and deployment on the ksqlDB server or cluster. The example demonstrates how to create a powerful UDF using an autoencoder neural network for real-time anomaly detection in car sensor data, showcasing its potential for continuous processing of large event streams. With this capability, users can leverage custom functions without requiring extensive programming knowledge, making it accessible to both data engineers and scientists.