HR GPT - A Comprehensive Tool for HR Executives

Welcome to HR GPT, a one-stop tool designed for HR executives to streamline the hiring process. This tool offers capabilities for candidate screening, background verification, finding better fitting job roles, and checking average salary ranges according to market standards.

Features

1. Candidate Screening:

   • Extracts tokens from resumes and job descriptions using Retrieval-Augmented Generation (RAG).
   • Creates embeddings with NVIDIAEmbeddings.
   • Stores embeddings using FAISS for efficient similarity search.

2. Background Verification:

   • Uses a LangChain Agent to perform web searches for background verification.
   • Calculates the candidate's age based on available information.

3. Job Fit Analysis:

   • Performs web searches to find job roles better suited to the candidate's skills.
   • Can be tailored to search for job openings in specific companies.

4. Salary Range Check:

   • Searches the web for average salary ranges for the candidate's skills and job role.

Technologies Used

• NVIDIA NIM Endpoints for high-performance inference.
• NVIDIA NeMo™ Guardrails to control LLM output.
• Streamlit for an interactive web interface.
• Plotly for data visualization.
• FAISS for fast similarity search.
• LangChain for building LLM applications.
• PyPDF2 for PDF handling.

Installation

1. Clone the repository:

   git clone - https://github.com/rutujakokate430/NvidiaGenAIAgentContest/tree/main
   cd hr-gpt

2. Install the required packages:

   pip install -r requirements.txt

3. Set up your API keys in the key.py file:

   nvidia_api_key = "your_nvidia_api_key"
   serp_api_key = "your_serp_api_key"

4. Run the application:

   streamlit run app.py

Configuration

• The guardrails configuration is located in the ./config directory.
• Update the job description and experience level via the Streamlit sidebar.
• Incase of any issues, use dash.py with key.py for running the application without NVIDIA NeMo™ Guardrails.
• key.py stores the api keys which are then called in dash.py

Usage

1. Upload the candidate's resume in PDF format.
2. Enter the job description in the provided text area.
3. Select the experience level from the dropdown menu.
4. View the results for candidate screening, background verification, better fitting roles, and salary range.

Acknowledgments

• Thanks to NVIDIA for the contest and the high-performance NVIDIA NIM Endpoints for inference.
• Special appreciation for NVIDIA NeMo™ Guardrails for ensuring accurate and controlled LLM outputs.