HR Resource Query Chatbot

---

An intelligent HR assistant chatbot that helps you query employee data in natural language.
It combines semantic retrieval with LLM-powered generation using a hybrid RAG (Retrieval-Augmented Generation) approach.

With this tool, you can ask:

• "Show me backend developers in Bengaluru with 5+ years of experience"
• "Who are the UI/UX designers skilled in Figma?"
• "List employees with leadership roles in sales"

and get chat-style responses with matching employee details.

---

Features

• Semantic Search → Search employees by role, skills, experience, and location.
• Conversational Chat → Ask questions in plain English.
• FastAPI Backend → Exposes REST endpoints for integration.
• Streamlit Frontend → Interactive chat-like interface.
• Hybrid RAG Engine → Combines embeddings (sentence-transformers) with Gemini LLM for contextual answers.
• Extendable Dataset → Add/update employees in data.py or link to your HR database.
• Fallback Mode → Even without Gemini API, retrieval still works.

---

System Architecture

---

Project Structure

HR_ChatBot/
│── 📜 app.py             # Streamlit frontend (chat interface)
│── 📜 main.py            # FastAPI backend (API endpoints)
│── 📜 rag.py             # RAG engine (semantic search + LLM)
│── 📜 data.py            # Employee seed dataset
│── 📜 requirements.txt   # Python dependencies
│── 📜 README.md          # Project documentation
│
└── 📂 assets/            # (Optional) Screenshots & demo images
    ├── ui_demo.png
    └── architecture.png

---

Setup & Installation

1. Clone Repository

git clone https://github.com/your-username/HR_ChatBot.git
cd HR_ChatBot

2. Create Virtual Environment

python -m venv venv

# Windows
venv\Scripts\activate

# macOS/Linux
source venv/bin/activate

3. Install Dependencies

pip install -r requirements.txt

---

Configure Gemini API

Obtain an API key from Google AI Studio. Set it as an environment variable:

# Windows PowerShell
$env:GEMINI_API_KEY="YOUR_KEY"

# macOS/Linux
export GEMINI_API_KEY="YOUR_KEY"

If no key is set, the chatbot falls back to retrieval-only answers.

---

Run the App

Start Backend (FastAPI)

uvicorn main:app --reload

• Open API docs → http://localhost:8000/docs

Start Frontend (Streamlit)

streamlit run app.py

• UI runs at → http://localhost:8501 -Update Backend URL in the UI if backend is running elsewhere.

---

Screenshots

Chat UI (Streamlit)

Employee Search Result

---

Employee Dataset

• Seed data is in data.py.
• Each employee contains:

{
  "id": 1,
  "name": "Alice Johnson",
  "role": "Backend Developer",
  "skills": ["Python", "Django", "PostgreSQL"],
  "experience": 5,
  "location": "Bengaluru"
}

• Extendable: Replace with CSV, DB, or API integration.

---

Notes

• Works with Python 3.10 – 3.13.
• For Python 3.13, use PyTorch ≥ 2.6.0.
• For GPU users, install CUDA-enabled PyTorch manually:

# Example for CUDA 12.1 on Windows
pip install torch --index-url https://download.pytorch.org/whl/cu121

• If Gemini API is down, app gracefully falls back to retrieval answers.

---

Contributing

Want to improve this project? Contributions are welcome!

• Fork the repo
• Create your feature branch → git checkout -b feature/my-feature
• Commit changes → git commit -m "Add my feature"
• Push branch → git push origin feature/my-feature
• Open a Pull Request

---

Author

Sourav Sharma Developer. Maker. Privacy-first AI enthusiast. Find me on GitHub → @Sourav-x-3202

---

Star This Project

If you found this useful, helpful, or inspiring — please consider starring the repository. It helps others discover the project and keeps development going

---

License

This project is licensed under the MIT License. © 2025 Sourav Sharma