
aws-samples/rag-with-amazon-bedrock-and-pgvector
📦 Open Source Projectaws-samples
A production-ready reference architecture for deploying RAG applications using Amazon Bedrock and PGVector on RDS.
The rag-with-amazon-bedrock-and-pgvector project is a robust reference implementation designed to bridge the gap between AI prototyping and production deployment. It utilizes a modern stack including Python, LangChain, and AWS CDK to orchestrate a complete RAG pipeline. Key technical components include Amazon Bedrock for accessing foundation models, and Amazon RDS for PostgreSQL with the pgvector extension, which allows for high-performance vector embeddings storage and retrieval.
The architecture is modular and secure, incorporating AWS Cognito for user management, Secrets Manager for credential handling, and ECS Fargate for containerized application hosting. By providing a structured approach to infrastructure, the repository allows developers to focus on application logic rather than cloud plumbing. It demonstrates best practices for managing vector databases in a cloud-native environment, handling document ingestion, and integrating LLMs into a web-based interface. This sample is particularly valuable for teams looking to standardize their RAG deployment patterns on AWS, ensuring scalability, maintainability, and security from the outset.
💡Highlights
- ├─Uses Amazon RDS with pgvector
- ├─Full AWS CDK infrastructure code
- └─Integrated Cognito authentication
🎯For
- ├─Cloud Architects
- ├─AI Engineers
- └─Full-stack Developers