A developer-focused RAG system that ingests technical documentation, architecture diagrams, code repositories, and incident reports to provide contextual answers, accelerate onboarding, and reduce troubleshooting time across engineering teams.
┌─────────────────────────────────────────────────────────────┐
│ Engineering Intelligence Hub │
│ │
│ Frontend (Dark dev UI · Highlight.js · Markdown) │
│ │ REST │ │
│ ┌────▼────────────────────────────────────────────────┐ │
│ │ FastAPI Backend │ │
│ │ POST /upload → processor.py → chunks │ │
│ │ (.py .js .ts .md .yaml .json .pdf .go .rs …) │ │
│ │ │ │ │
│ │ sentence-transformers │ │
│ │ (all-MiniLM-L6-v2, FREE) │ │
│ │ │ │ │
│ │ ChromaDB + filetype metadata │ │
│ │ (cosine similarity) │ │
│ │ │ │ │
│ │ POST /query → top-k retrieval → LLM │ │
│ │ (GPT-4o-mini / flan-t5) │ │
│ │ │ │ │
│ │ Output: Answer · Code Example · Architecture Notes │ │
│ └─────────────────────────────────────────────────────┘ │
└─────────────────────────────────────────────────────────────┘
- Ingests 14 file types:
.py .js .ts .go .rs .java .md .yaml .yml .json .sh .html .txt .pdf - Filetype metadata stored per chunk — retrieval is file-type aware
- Structured AI output:
- Answer — precise technical explanation
- Code Example — syntax-highlighted code blocks
- Architecture Notes — design decisions, trade-offs
- Related Resources — specific docs/modules from knowledge base
- Full syntax highlighting via Highlight.js (GitHub Dark theme)
- Persistent ChromaDB vector store
cd backend
pip install -r requirements.txt
cp ../.env.example .env # Add OPENAI_API_KEY (or skip for local HuggingFace)
uvicorn main:app --reload --port 8000
# Open frontend/index.html in browser| Scenario | Example Query |
|---|---|
| Onboarding | "How does the authentication flow work?" |
| Debugging | "Why does the payment service timeout under load?" |
| Architecture | "What's the data flow between the API gateway and microservices?" |
| Incident review | "What caused the P0 outage on 2024-03-15?" |
| Code review | "What are the patterns used in this codebase?" |
engineering-intelligence-hub/
├── backend/
│ ├── config.py # Engineering domain settings + 14 supported filetypes
│ ├── processor.py # Multi-format loader + chunker (stores filetype metadata)
│ ├── pipeline.py # ChromaDB + embeddings + LLM generation
│ ├── main.py # FastAPI app
│ └── requirements.txt
├── frontend/
│ └── index.html # Dark dev UI with syntax highlighting (no build step)
├── .env.example
└── README.md
FastAPI · sentence-transformers · ChromaDB · OpenAI GPT-4o-mini · pypdf
Vanilla JS + marked.js + Highlight.js (no npm, no build step)