Building AI Powered Applications with Java
summary
Shows how to build RAG pipelines with Spring AI and LangChain4j connecting to OpenAI (GPT‑4o/4‑mini) for chat, but warns about real issues: wrong parser breaks ASCIIDoc/Markdown snippets, embedding-dimension mismatches with the model, and high latency if chunking/embeddings run at query time. Demo loads 397 docs into memory, creates ~2,800 chunks, and stores embeddings into a vector store using Oracle vector support via testcontainers — it works for basic Q&A but yields non-deterministic and sometimes incorrect answers.
AI-written summary · about
chapters (13)
- 0:00Introduction and welcome
- 1:50What is AI and its evolution
- 3:38Why use AI in development
- 8:15When to start AI projects
- 8:37Spring AI vs LangChain4j
- 12:52RAG and private data handling
- 17:47Chunking and embeddings best practices
- 21:59Demo setup and first RAG example
- 28:08Fixing parser and document formats
- 31:25Testing LLM outputs and non-determinism
- 35:19Vector database integration demo
- 41:27Future of AI tools and MCP
- 43:34AI impact and resources summary
related talks
01h 03mBuilding AI powered Applications with Java
42mBuilding AI Powered Applications with Java
55mAI in Java? Meet LangChain4j and Build Intelligent Services
46mGenerative AI and Java
53mUnleash the Power of AI Agents with Java Building Intelligent Application
28mAPIs and GenAI : Your Perfect Combo with LLM-Powered Frameworks.
02h 49mJava meets AI: Build LLM-Powered Apps with LangChain4j
41mFrom Chat to RAG to MCP: Enhancing Java Applications with AI