An AI-powered resume chatbot that lets you upload resumes and ask natural language questions about candidates.
Run locally at http://localhost:5050.
- Upload Multiple Resumes - Supports PDF files, handles 20+ resumes
- Natural Language Queries - Ask questions in plain English
- Smart Search - Hybrid semantic + keyword search
- Candidate Comparison - Compare candidates side by side
- Skill Matching - Find candidates with specific skills
- Conversation Memory - Understands follow-up questions and pronouns
- Resume Validation - Automatically rejects non-resume documents
- "Whose resumes do you have?"
- "Who has Python experience?"
- "Tell me about John's background"
- "Compare the candidates for a backend role"
- "Who would be best for a data science position?"
- "What was she doing in Jan 2025?" (uses conversation context)
- "Who has AWS and React experience?"
| Component | Technology |
|---|---|
| Frontend | React + Tailwind CSS |
| Backend | FastAPI |
| LLM | OpenAI GPT-4o |
| Vector Database | ChromaDB |
| PDF Processing | PyPDF2 |
| Language | Python 3.11.9 |
┌─────────────────┐ ┌─────────────────┐ ┌─────────────────┐
│ │ │ │ │ │
│ PDF Upload │────▶│ Text Chunks │────▶│ ChromaDB │
│ │ │ + Metadata │ │ (Vectors) │
└─────────────────┘ └─────────────────┘ └─────────────────┘
│
▼
┌─────────────────┐ ┌─────────────────┐ ┌─────────────────┐
│ │ │ │ │ │
│ Response │◀────│ GPT-4o │◀────│ Hybrid Search │
│ │ │ │ │ (Semantic+KW) │
└─────────────────┘ └─────────────────┘ └─────────────────┘
- Python 3.11.9 or higher
- OpenAI API key
- Clone the repository
git clone https://github.com/yourusername/resume-rag.git
cd resume-rag- Create virtual environment
python -m venv venv
source venv/bin/activate # On Windows: venv\Scripts\activate- Install dependencies
pip install -r requirements.txt- Set up environment variables
# Create .env file
cat > .env <<EOF
OPENAI_API_KEY=your_api_key_here
TAVILY_API_KEY=your_tavily_api_key_here
EOF- Run the application
python3 resume_rag.pyThe app will open at http://localhost:5050
resume-rag/
├── resume_rag.py # FastAPI backend + chat logic
├── templates/
│ └── index.html # Tailwind + React shell
├── static/
│ └── react_app.js # React UI logic
├── requirements.txt # Python dependencies
├── .env # Environment variables (not in repo)
├── .gitignore # Git ignore file
├── README.md # This file
├── chroma_db/ # Vector database storage (auto-created)
| Setting | Default | Description |
|---|---|---|
| Chunk Size | 300 words | Size of text chunks for processing |
| Chunk Overlap | 75 words | Overlap between consecutive chunks |
| Max Chunks to LLM | 70 | Maximum chunks sent to GPT-4o |
| Conversation History | 5 | Number of previous Q&A pairs remembered |
