Coffee + Software

A live coding walkthrough of a Spring AI workshop covering four progressive labs: basic LLM chat with Spring Boot, RAG (Retrieval Augmented Generation) using Redis as a vector store, and agentic workflows using the Embabel framework's goal-oriented action planning. The workshop runs entirely locally using Ollama with the Qwen 2.5 Coder model (under 1GB) and Redis in Docker, requiring only 8GB RAM. Key concepts covered include system prompts, LLM hallucinations and their cutoff dates, document ingestion and chunking, vector similarity search, and type-safe agentic pipelines using Java records. Spring AI 2.0.0-M4 and Embabel 0.3.4 are used, though they are not yet compatible with each other.

Spring AI 2.0.0-M4 and Embabel 0.3.4 📱