Sprint 1: RAG MVP Foundation

Goal

Build the smallest useful platform slice: FastAPI, Docker services, policy ingestion, sample SEC ingestion, citation-shaped RAG answers, and a React dashboard.

Why This Sprint Matters

The project starts with retrieval and evidence before agent orchestration. In a financial setting, unsupported answers are more damaging than limited answers. Sprint 1 proves the platform can ingest documents, retrieve evidence, and show sources.

What Was Built

• FastAPI backend with health, config, ingestion, and chat endpoints
• Docker Compose for PostgreSQL, Qdrant, and Redis
• Local policy documents for enterprise governance examples
• Basic document chunking and citation-shaped responses
• React dashboard with chat, sources, trace, metrics, and system status

Architecture / Workflow

flowchart LR
    UI[React Dashboard] --> API[FastAPI]
    API --> Ingest[Policy / SEC Sample Ingestion]
    Ingest --> Store[Local + Vector Store]
    API --> Chat[Chat Endpoint]
    Chat --> Retrieve[Retrieve Evidence]
    Retrieve --> Answer[Cited Answer]

Key Files And APIs

• backend/app/api/routes.py
• backend/app/services/ingestion_service.py
• backend/app/services/rag_service.py
• POST /api/ingest/policy
• POST /api/ingest/sec
• POST /api/chat

Validation Commands

.\.venv\Scripts\python -m pytest
cd frontend
npm run build

Demo Talking Points

Explain that the MVP intentionally avoids complex agents until retrieval and citations are stable. The first credibility signal is evidence-backed answering.

What Changed From Previous Sprint

This is the foundation sprint. It establishes the repo structure, running services, backend API, and dashboard.