Project Babylon enables developers to build and run AI models - such as LLMs, image classifiers, or object detection algorithms - directly in Java. With Code Reflection, machine learning logic can be defined in plain Java code, eliminating the need for Python or external model files. By leveraging the Foreign Function and Memory (FFM) API, Babylon can connect Java code to native runtimes like ONNX for fast, GPU-accelerated inference. Additionally, the Heterogeneous Accelerator Toolkit (HAT) enables developers to write and compose compute kernels in Java, making it easy for Java libraries to tap into GPU power for high-performance computing.

This session introduces Babylon’s upcoming features and demonstrates how you can integrate AI capabilities into the Java ecosystem, appealing to both library maintainers and developers looking to incorporate AI into their Java applications.

Recorded at Devoxx Belgium 2025.