HR RAG System

Overview

HR RAG System is a Retrieval-Augmented Generation (RAG) application for querying HR employee data using natural language. The system transforms structured employee data into searchable documents, retrieves relevant context using embeddings, and generates answers using an LLM.

Architecture

HR Data → Document Processing → Embeddings (OpenAI) → Chroma Vector Store → Retrieval → LLM → FastAPI → Response

Features

• Converts HR employee data into structured documents
• Generates embeddings using OpenAI
• Stores vectors in Chroma for efficient retrieval
• Uses LangChain for retrieval and response generation
• Exposes a FastAPI endpoint for querying the system
• Returns context-aware, generated answers from HR data

Tech Stack

• Python
• OpenAI
• LangChain
• Chroma (Vector Database)
• FastAPI

Example Use Cases

• Summarize an employee profile
• Query HR data using natural language
• Retrieve insights from employee records
• Generate answers grounded in internal data

API Usage

Endpoint

POST /query

Request Body

{
  "question": "Give me a summary of employee 1"
}

Response

{
  "answer": "..."
}

Deployment

This project is structured as an API-based AI system using FastAPI. It can be deployed to cloud platforms, but is currently provided as a local or development environment due to dataset size and deployment constraints.

Project Structure

hr-rag-system/
├── hr_employee_rag.py        # RAG pipeline and core logic
├── hr_rag_api.py             # FastAPI application
├── hr_data/
│   └── employees/            # Processed employee documents
├── HR-Employee-Attrition.csv # Source dataset
├── requirements.txt          # Dependencies
└── README.md

Getting Started

Clone the repository:

git clone https://github.com/moatazsaad/hr-rag-system.git
cd hr-rag-system

Create a virtual environment:

python -m venv env

Activate environment:

Windows:

env\Scripts\activate

macOS/Linux:

source env/bin/activate

Install dependencies:

pip install -r requirements.txt

Set your OpenAI API key in a .env file:

OPENAI_API_KEY=your_api_key_here

Run the API:

uvicorn hr_rag_api:app --reload

Example Query

curl -X POST "http://127.0.0.1:8000/query" \
-H "Content-Type: application/json" \
-d '{"question": "Give me a summary of employee 1"}'

Project Highlights

• End-to-end RAG pipeline from structured HR data to generated answers
• Combines embeddings, vector database, and LLMs
• Exposes a production-style API using FastAPI
• Demonstrates applied GenAI and retrieval-based systems