Project 09: Final Enterprise AI System

Objective

Build a production-style AI system that combines document intelligence, RAG, agents, evaluation, cloud architecture, security, operations, and product justification.

Required Features

• Authentication.
• Role-based access control.
• Document ingestion.
• RAG with citations.
• Tool calling.
• Safe write workflows with confirmation.
• Audit logs.
• Evaluation pipeline.
• Monitoring and cost tracking.
• Docker deployment.
• Kubernetes deployment option.
• Cloud architecture documentation.
• ADRs.

Milestones

• Milestone 1: Requirements and architecture.
• Milestone 2: Backend and data model.
• Milestone 3: RAG ingestion and retrieval.
• Milestone 4: Agent tools and authorization.
• Milestone 5: Evaluation and regression testing.
• Milestone 6: Security hardening.
• Milestone 7: Deployment and operations.
• Milestone 8: Final documentation and demo.

Final Demo Script

• Upload or ingest documents.
• Ask a question and receive cited answers.
• Attempt an unauthorized action and show denial.
• Run an authorized read tool.
• Preview a write action.
• Confirm the write.
• Show audit log.
• Run evaluation report.
• Show architecture and cost estimate.

Tips

• Keep the demo path simple and reliable.
• Prefer fewer features with stronger engineering quality.
• Track every major decision in ADRs.
• Include known limitations honestly.