AI-Driven RAG Chatbot & Test Automation

# AI Platform – RAG Chatbot + AI Test Automation

Looking for an experienced **AI/Python developer** to build an MVP consisting of two related AI modules:

### 1. RAG Chatbot

Build a secure chatbot that can answer questions using uploaded/internal knowledge.

Key expectations:

- Upload and process **PDF, DOCX, text and other documents**
- Document chunking, embeddings, vector search & retrieval
- Source-grounded answers with **citations/references**
- Conversation history, context & memory management
- Guardrails, hallucination control and prompt-injection protection
- RAG evaluation: **relevance, groundedness/faithfulness, retrieval quality, hallucination, etc.**
- Support multiple LLM providers/models
- Token usage & **cost tracking**

### 2. AI Test Generation & Automation

Core flow:

**PRD / BDD / PDF / Text / Jira → RAG → AI Test Cases → Review → Excel → Playwright → Execution → Dashboard**

Key expectations:

- Generate structured manual test cases with **requirement traceability**
- Review/edit and export test cases to **Excel**
- Generate **Playwright automation**
- Controlled test execution
- Pass/Fail/Skipped results
- Current + **historical dashboards**, trends and coverage

### Shared AI / Platform Capabilities

The architecture should support:

- **Multi-agent orchestration** – Planner, Generator, Executor, Evaluator, etc.
- Agent/tool communication and **MCP integration** where appropriate
- RAG & context management
- Memory & caching
- AI evaluations
- Guardrails & security
- LLM/model abstraction
- Logging & observability
- Retry/error/timeout handling
- **Infinite-loop protection**
- Token/context/cost dashboard
- Authentication & secure document handling
- Docker/cloud deployment

Likely technologies: **Python, FastAPI, React/Next.js, Playwright, PostgreSQL/pgvector, Docker, MCP, OpenAI/Anthropic/Gemini, DeepEval/RAGAS**, or suitable alternatives.

### When Applying

Please provide:

- Relevant **RAG / Agentic AI / Playwright / LLM evaluation** experience
- Examples/links to similar projects
- Proposed architecture & technology stack
- What you would build vs. use off-the-shelf
- **Milestone-wise timeline & cost**
- Expected hosting/LLM running costs

We prefer **simple, practical and extensible architecture over unnecessary complexity**.

Please start your proposal with **“AI-QE-RAG”** so we know you've read the requirement.