Quanta Live: Building Local AI Search with pgvector & LangChain4j - Part 2

Description

Building Quanta (Part 2) – Vector Search with Quarkus, pgvector & Ollama In this session, we’re diving deep into the brains of Quanta — the part that makes it actually understand your files. 🧠 We’ll take our local AI file searcher from Part 1 and wire up vectorization and semantic search using Quarkus, LangChain4j, Ollama, and PostgreSQL with pgvector. You’ll see how to: - Extract text from files with Apache Tika - Generate embeddings using Ollama (LLaMA 3.2) - Store and search them efficiently with pgvector - Query your files in natural language: no more hunting through folders! It’s a full-stack AI search pipeline, built live and explained step by step. Expect real coding, debugging, and a few “oh-that’s-why-it-wasn’t-working” moments along the way. 😅 By the end, you’ll understand: ✅ How embeddings work in practice ✅ How to integrate pgvector with Quarkus + LangChain4j ✅ How to perform similarity searches that actually return meaningful results 🧱 Tech Stack Quarkus · LangChain4j · Ollama (LLaMA 3.2) · PostgreSQL + pgvector · Apache Tika · Docker · Next.js 📦 Source Code (Live) 👉 https://github.com/JohannesRabauer/quanta 📅 Timestamps (coming soon) 💬 Drop your questions in chat, join the live coding fun, and help shape how Quanta evolves. #Quarkus #LangChain4j #Ollama #pgvector #AIFileSearch #LocalAI #Postgres #LLaMA3 #VectorSearch #LiveCoding #AI #Nextjs