For years, Python has dominated the data science landscape, leaving Java developers to bridge the gap between model experimentation and enterprise production. This session breaks that barrier by demonstrating how to build, train, and deploy high-performance Machine Learning models entirely within the Java ecosystem. We will explore the Tribuo library’s robust architecture, highlighting its unique ability to provide compile-time type safety for ML models.

We will then dive into the modern developer's workbench: Jupyter Notebooks for Java. We start from a clean environment and walk through the exact setup: installing Jupyter + the IJava kernel, connecting Tribuo to Oracle AI Database 26ai via JDBC, loading data directly from tables (Iris example used in the live demo), creating Tribuo data sources, splitting datasets, training models, and evaluating accuracy and confusion matrices — all inside live notebook cells.

Furthermore, the session will bridge the gap between experimentation and enterprise-grade deployment by demonstrating how to utilize Oracle AI Database 26ai for database-centric ML architecture, effectively bringing the logic to the data to minimize latency and streamline the path to production.