# GitHub Presentation

Use these snippets for the GitHub repository description, About section, or pinned project notes.

## Repository Description

AI equity research agent with resilient workflow orchestration, Redis Lua single-flight, pgvector RAG, versioned reports, evidence tracing, and RAG evaluation.

## Short Pitch

FinSight is an open-source AI equity research agent backend. It combines Spring Boot, PostgreSQL/pgvector, Redis, RabbitMQ, and a FastAPI AI sidecar to generate source-grounded stock research reports with recoverable agent workflows, trustworthy cache versioning, and measurable RAG quality.

## Project Tags

```text
ai-agent
rag
spring-boot
postgresql
pgvector
redis
rabbitmq
financial-research
workflow-orchestration
llm-evaluation
```

## README Badge Ideas

```markdown
![Java](https://img.shields.io/badge/Java-17-blue)
![Spring Boot](https://img.shields.io/badge/Spring%20Boot-3.3-green)
![PostgreSQL](https://img.shields.io/badge/PostgreSQL-pgvector-blue)
![Redis](https://img.shields.io/badge/Redis-single--flight-red)
![RabbitMQ](https://img.shields.io/badge/RabbitMQ-workflows-orange)
```

## Social Post Draft

I built FinSight, an AI equity research agent backend that focuses on the infrastructure around LLM output:

- recoverable agent workflow state machine;
- Redis Lua single-flight and fencing tokens;
- PostgreSQL/pgvector hybrid retrieval;
- report versioning with `dataSnapshotHash`;
- evidence-grounded RAG answers;
- hallucination risk and consistency evaluation.

It is a backend-heavy AI Agent project rather than a thin chat wrapper.