Enterprise-grade Agentic AI system for financial transaction investigation, compliance-aware reasoning, and production-ready LLM tooling.
This project demonstrates how to build regulated, auditable, and scalable agentic AI systems similar to those used in large fintech environments.
- Agent orchestration layer
- OpenAI structured function calling
- pgvector-powered RAG (Postgres native vectors)
- Policy document embedding + retrieval
- Tool execution framework
- Repository pattern + service layer
- Dockerized infrastructure
- Production-style separation of concerns
- OpenAPI docs via FastAPI