HR-LLM-Assistant

An intelligent HR assistant built using Large Language Models (LLMs) to help HR managers and employees interact seamlessly. This chatbot can handle HR-related queries such as leave policies, benefits, and onboarding procedures while analyzing employee engagement metrics.

Features

• HR assistant chatbot answering queries related to HR policies.
• Retrieval-Augmented Generation (RAG) to provide context-based answers using HR documents.
• Engagement metrics like average question and response sentiment.
• Support for HR managers to manage leave policies, benefits, and onboarding processes.

Installation

Step 1: Clone the Repository

Clone the repository to your local machine:

git clone https://github.com/Madhushan13/HR-LLM-Assistant.git
cd HR-LLM-Assistant

Step 2: Set Up the Environment

Create a Virtual Environment:

To create a virtual environment, run the following command:

python -m venv venv

Activate the Virtual Environment:

On Windows, run:

venv\Scripts\activate

On Mac/Linux, run:

source venv/bin/activate

Step 3: Install Dependencies

Install the required Python dependencies using pip:

pip install -r requirements.txt

Step 4: Create a .env File

To keep sensitive data like your OpenAI API key, create a .env file in the root of your project directory.

In the root folder (where app.py is located), create a file named .env.

Add the following content, replacing your_openai_api_key with your actual OpenAI API key:

OPENAI_API_KEY=your_openai_api_key

Step 5: Run the Flask App

Once everything is set up, you can run the chatbot with the following command:

python app.py

Your chatbot should now be running locally at http://localhost:5000.

Step 6: Testing the Chatbot with Postman

Once your bot is running, you can test it using Postman.

1: Test the /ask Endpoint: Open Postman and create a POST request.

Use the following URL:

http://localhost:5000/ask

In the Headers section, add:

• Key: Content-Type

• Value: application/json

In the Body section, select raw and JSON format. Then add the following JSON to simulate a user query:

{
    "query": "What is the leave policy?"
}

Click Send. You should receive a response like this:

{
    "response": "Employees are entitled to 10 days of paid leave per year."
}

2: Test the /analyze-engagement Endpoint: Create a GET request.

Use the following URL:

http://localhost:5000/analyze-engagement

Click Send. You should receive a response like:

{
    "avg_question_sentiment": 0.5,
    "avg_response_sentiment": 0.6
}

Technologies Used

• Python - For programing language
• Flask - For backend API
• OpenAI GPT-3 - For natural language processing
• Langchain - For LLM pipeline management
• ChromaDB - For storing document embeddings
• TextBlob - For sentiment analysis
• Pinecone - For vector database storage (optional)