Hybrid Semantic web with MongoDB and Vector Search using LangChain.js and Phi-3

This talk provides the way to build intelligent applications powered by semantic search and GenAI over any type of data using a full-featured vector database, LLM framework, and open small language models with low-cost development. For the frontend site, I will use Vercel using Vue.js. For the backend site, I will use Node.js for MongoDB API and Ollama API. The agenda is described below. - Preparing local environments - Ollama with Phi-3-mini - Cloudflare Tunnel - MongoDB Atlas and Atlas Vector Search using Podman - Integrate MongoDB and Ollama using Ollama JavaScript library - Hybrid Semantic search with Reciprocal Rank Fusion (RRF) - Deploying GenAI web application with a low-cost budget to Vercel (Vue.js) and MongoDB App Services (Node.js) - Key takeaways and challenges.