Bring Your Own Data! Build Better AI Applications
summary
Demo uses Weaviate and JavaScript to wire a vector database into a retrieval pipeline so an LLM answers only from owned docs. It builds embeddings, runs similarity search over a 500‑song corpus and a Rolling Stone top‑50 set, then augments prompts for generation. The demo surfaces the exact context used, shows the new client only supports nodejs, and demonstrates RAG with retrieval‑augmented generation while recovering from a few live demo hiccups.
AI-written summary · about
chapters (9)
- 0:00Introduction
- 1:56Why your data matters
- 8:19LLM problems and hallucinations
- 12:11RAG for open-book AI
- 17:09Live demo: RAG with Verba
- 26:05RAG security and context
- 32:51Vector databases in RAG
- 35:02Semantic search in JavaScript
- 43:08Summary
related talks
31mBring AI based Search to Your Website
08mRethinking your data stack: The future of AI apps and vector databases
48mGoing Beyond Prompt Engineering
36mTypescript, You and AI Search
34mJavaScript-Powered AI: Crafting Intelligent Systems with LLMs and Vector Databases
23mWebAI: Building the next user experience in the browser
14mBring AI to Your Data
01h 17mGraph Database and Retrieval Augmented Generation