Embabel Explained: Building Production-Ready AI Agents on the JVM

In this CyberJAR livestream episode, Catherine interviews Dr. Rod Johnson, the founder of the Spring Framework, about his new project, Embabel, an agentic framework for the JVM. Rod explains why the JVM is often dismissed in AI discussions, emphasizing the immaturity of Gen AI in enterprise settings and the dominance of Python in data science. He argues that Java frameworks are now superior and that JVM-based systems are ideal for building Gen AI applications within enterprises. Rod details Embabel's innovative approach to orchestration using goals, actions, and planning, contrasting it with state machines and sequential steps found in Python frameworks. He also touches on the relationship between Embabel and Spring AI, its deterministic nature, and its capabilities in handling tools and RAG (Retrieval Augmented Generation). The conversation includes live demos of Embabel, showcasing its features, type safety, and potential for use in various projects, from simple AI enhancements to complex workflows and chatbots.