Building generative AI–enriched applications with AWS & MongoDB Atlas
summary
MongoDB Atlas demonstrates a practical RAG pipeline using Atlas Vector Search plus integrations with LlamaIndex, LangChain and Semantic Kernel, exposing a $vectorSearch stage, numCandidates tuning and an HNSW index to trade recall for millisecond latency (authors recommend ~10–20:1 candidates:limit). Change streams keep indexes synced, Atlas manages hosting across 100+ regions, and dedicated Search Nodes (2–32, memory-optimized) separate search from OLTP so large vector indexes fit in memory and perform at scale.
AI-written summary · about
chapters (7)
- 0:00Introduction to vector search
- 1:14Fundamentals of vectors
- 8:07Evolution of embedding models
- 9:36Implementing Atlas search
- 14:24Retrieval augmented generation
- 17:44Architecture and challenges
- 21:03Scaling with search nodes
related talks
22mGenerative AI application stack and providing long term memory to LLMs
25mRedefining Search: Vector Search with OpenAI and MongoDB
36mUnleashing the Power of Vector Databases for AI-Driven Applications
50mHow to design and build AI applications with vector search using Azure OpenAI & Azure Cosmos DB
12mBuilding Generative AI Apps with Azure Cosmos DB
40mBuilding Intelligent Search Systems With Agentic RAG and OpenSearch
47mRAG at scale: production-ready GenAI apps with Azure AI Search
46mBuild an in-place Amazon S3 data lake with