Skip to main content

Building RAG Applications with MongoDB Atlas

Welcome to the MongoDB-RAG Workshop! Throughout this hands-on experience, you'll learn how to build Retrieval-Augmented Generation (RAG) applications using MongoDB Atlas Vector Search and the mongodb-rag library.

What You'll Learn​

In this workshop, you'll:

  1. Understand RAG fundamentals - Learn how vector search and LLMs work together
  2. Set up MongoDB Atlas - Create and configure an Atlas cluster with Vector Search
  3. Generate embeddings - Transform documents into vector embeddings
  4. Build a complete RAG application - Develop a full application with document ingestion and search
  5. Implement advanced techniques - Explore hybrid search, metadata filtering, and chunking strategies
  6. Deploy to production - Learn best practices for scaling your RAG system

Prerequisites​

To get the most out of this workshop, you should have:

  • Basic knowledge of JavaScript/Node.js
  • Familiarity with MongoDB concepts (collections, documents, queries)
  • A MongoDB Atlas account (free tier is sufficient)
  • An OpenAI API key (or Ollama installed locally)
  • Node.js (v16+) installed on your system

Workshop Structure​

This workshop is designed to be completed in sequence, with each section building on the previous ones:

  1. RAG Concepts - Understanding the fundamental concepts
  2. Setting up MongoDB - Configuring your Atlas environment
  3. Creating Embeddings - Generating and storing vector embeddings
  4. Building a RAG App - Creating a full application
  5. Advanced Techniques - Improving relevance and performance
  6. Production Deployment - Scaling and monitoring considerations

Estimated Time​

The complete workshop takes approximately 2-3 hours to complete. Each section is designed to be completed in 20-30 minutes.