Speakers

Covering how it is possible to incorporate cutting-edge open source AI capabilities into Java applications, and build full Retrieval Augmented Generation solutions without leaving the JVM ecosystem.  Looking at options for local model deployment, bindings to use popular open source frameworks with the JVM and the latest open source models.Open-source AI models and frameworks have experienced tremendous growth recently, and while the most popular interfaces are for the Python language (or C++ for working with the internals), Java Virtual Machine (JVM) bindings exist for some of the most important projects.This talk will delve into some of the tools and technologies available to integrate open-source AI models with JVM projects, including:Ollama SDK - This communicates with the Ollama server, which is intended for running models locally.The Java Native Interface (JNI) bindings for Llama.cpp, which is a popular open source project for running many of the most popular large language models locally using the CPU, GPU or Apple’s Neural Engine.TensorFlow for JavaDeep Java Library (DJL) from AmazonThe JNI (Java Native Interface) bindings for PyTorchJVector - A Vector Search Library, which can be used with the other technologies above to enable the building of fully open-source Retrieval Augmented Generation (RAG) solutions controlled from the JVM. Read More...