Multi-tenant vector search with Amazon Aurora PostgreSQL and Amazon Bedrock Knowledge Bases

This article discusses a fully managed approach to building a multi-tenant vector search solution using Amazon Aurora PostgreSQL and Amazon Bedrock Knowledge Bases, specifically for a home survey document management system.

• Architecture enables users to ask natural language questions about property survey documents
• Uses Amazon Bedrock Knowledge Bases to manage vector embedding ingestion and retrieval
• Implements multi-tenant data isolation through metadata filtering
• Allows retrieving contextually relevant document chunks for prompt augmentation
• Supports flexible embedding models like Amazon Titan Embeddings V2

Key benefits include simplified RAG workflow, tenant data isolation, and the ability to leverage fully managed services for vector search without complex custom integrations.