Repository files navigation

A comprehensive framework for building, deploying, and managing Agentic AI Retrieval-Augmented Generation (RAG) workflows on Google Cloud Platform, with a focus on medical customer support applications.

This framework provides a complete end-to-end solution for implementing production-grade RAG systems with agentic capabilities. It addresses the unique challenges of retrieval-augmented generation workflows, including vector database management, agent orchestration, document processing, compliance, evaluation, and cost optimization.

The framework is built on Google Cloud Platform and leverages key agent technologies:

• Agent Development Kit (ADK) for building multi-agent systems
• Google Agentspace for agent discovery and governance

• Vector Database Integration: Efficient storage, indexing, and retrieval of medical knowledge embeddings
• Agent Orchestration: Coordinating tools, reasoning chains, and stateful interactions for medical support
• Document Processing: Chunking, embedding, and indexing medical documents effectively
• Compliance Framework: HIPAA-compliant handling of sensitive medical information
• Real-time Serving: Low-latency, high-throughput serving with streaming responses
• Cost Optimization: Managing the significant costs of LLM API calls and embedding generation
• React-based UI: Modern, responsive interface for medical support agents
• Monitoring & Governance: Comprehensive dashboards for system oversight

The framework consists of several interconnected layers, each addressing specific aspects of RAG system development and deployment. See theArchitecture Documentation for detailed diagrams and explanations.

• Docker and Docker Compose
• Google Cloud SDK
• Node.js and npm (for UI development)
• Python 3.9+

1. Clone this repository:

   git clone https://github.com/yourusername/enhanced-rag-framework.gitcd enhanced-rag-framework

2. Set up environment variables:

   cp .env.example .env# Edit .env with your configuration

3. Start the services:

   docker-compose up

4. Access the UI athttp://localhost:3000

See theDeployment Guide for detailed instructions on deploying to Google Cloud Platform.

• Architecture
• Configuration Layer
• Execution Engine
• Validation Layer
• Governance Layer
• Infrastructure Components
• Developer Experience
• Deployment Guide
• CLI and SDK Usage
• Monitoring and Logging

This project is licensed under the MIT License - see theLICENSE file for details.

• Google Cloud Platform for providing the infrastructure and agent technologies
• The open-source community for their invaluable contributions