Pratik Patel is a Java Champion and lead developer advocate at Azul Systems. He wrote the first book on 'enterprise Java' in 1996, "Java Database Programming with JDBC" and “Developing Open Cloud Native Microservices”. An all around software and hardware enthusiast with experience in the travel, healthcare, telecom, financial services, and startup sectors. Helps to organize the Atlanta Java User Group, frequent speaker at tech events, and master builder of nachos.
The AI application space is moving rapidly and there are new & better LLM models coming out frequently. You’ve seen articles and videos of using Retrieval Augmented Generation (RAG) that look amazing - but once you start trying to use basic RAG to build your own applications, you suddenly hit a wall and things don’t work as expected… which is very frustrating.
In this session, we will take a step back, and understand the fundamentals of RAG. We’ll look at different RAG strategies and techniques. Understanding the fundamentals of LLMs is a good starting point, and then we’ll expand into the different approaches to RAG, and discuss how MCP (Model Context Protocol) can be used along with RAG. Here’s the topics we’ll cover:
- LLMs are semantic search based tools
- Vectorization and how it relates to semantic search
- Choosing the right tokenization strategy
- Embeddings
- Contextual Retrieval RAG
- Hybrid Search Enhance RAG
- MCP basics
- Difference between MCP and RAG
- Examples of the RAG techniques, MCP Server integration, and a complete application to give you concrete takeaways to build your own AI application
Searching for speaker images...