Azure Cosmos DB in AI Workflows: From Knowledge Retrieval to Microsoft Fabric Integrations
summary
Shows how Azure Cosmos DB stores embeddings and runs native vector search (embeddings from OpenAI models), performs hybrid search with a reciprocal rank fusion-style mix of vector plus full‑text, and uses the Cosmos SQL engine as a natural‑language→SQL endpoint (two-call flow: service URL + bearer token) to return top‑N records. Demonstrates Cosmos DB mirroring that converts snapshots and CDC streams into Delta/OneLake in near‑real‑time for Microsoft Fabric, then builds views and Power BI reports to surface analytics.
AI-written summary · about
chapters (8)
- 0:00Azure Cosmos DB in AI workflows
- 2:11Knowledge retrieval with RAG
- 6:24Vector search in Cosmos DB
- 8:51Natural language to SQL
- 11:00Fabric integrations with Cosmos DB
- 13:28Mirroring Cosmos DB in Fabric
- 14:31Power BI data transformation
- 16:17Summary
related talks
12mBuilding Generative AI Apps with Azure Cosmos DB
15mAzure Cosmos DB as Vector Database
50mHow to design and build AI applications with vector search using Azure OpenAI & Azure Cosmos DB
01h 16mAzure SQL ❤️ Open AI
03h 12mAzure Cosmos DB Conf 2024
40mUnleashing the Power of Vector Search with Azure SQL Database
50mBuilding AI ready applications
20mBuilding AI Applications with Azure Cosmos DB JavaScript SDK