Empowering AI Education with Rag: Building Domain-Specific LLM Agents for Enhanced...
summary
Shows a working pipeline that turns course notes and textbooks into embeddings, loads them into a vector DB, and wires that RAG layer to self‑hosted LLMs (Gemma 2B, Llama 27B/37B and a 30B fine‑tuned model) using Rust‑powered RAG components and WasmEdge/LamaEdge runtimes. Demo runs on an Intel Mac with 16 GB, copies a .wasm API server to an NVIDIA host, and reports slow first queries then faster followups — overall a portable, offline, self‑hosted teaching assistant.
AI-written summary · about
chapters (11)
- 0:00AI education with RAG
- 3:43AI-driven education challenges
- 7:01Domain-specific teaching assistants
- 10:17Self-hosted open-source LLMs
- 13:10WebAssembly for AI inference
- 17:54Demo: Lama Edge locally
- 21:49Demo: Portability on Nvidia GPU
- 26:21DevOps separation with WebAssembly
- 28:50Simplifying RAG with Gatet
- 31:00Demo: Book-powered RAG chatbot
- 36:02Open-source contributions
related talks
41mFrom RAGs to riches: Build an AI document interrogation app in 30 mins
23mWebAI: Building the next user experience in the browser
01h 34mHow to Construct and Assess Your First LLM Based RAG System
48mGoing Beyond Prompt Engineering
18mLLMs and GenAI: A Guide to Building Your Own RAG Model for Document Interaction
48mCrafting your own RAG system: Leveraging 30+ LLMs for enhanced performance
48mCrafting your own RAG system: Leveraging 30+ LLMs for enhanced performance
32mScaling Generative AI: Building Production-Ready LLM Applications