This project is an AI-powered Resume Matcher designed to help hiring managers efficiently find the best candidate matches for job openings. It leverages advanced language models and embeddings to analyze resumes and match them against job requirements.
- PDF Resume Processing: Extracts and processes resumes in PDF format.
- Semantic Embeddings: Uses Hugging Face models to create semantic embeddings for resumes.
- Vector Database: Stores and retrieves resumes using ChromaDB for efficient similarity searches.
- Conversational AI: Employs a conversational chain to provide intelligent responses for resume matching queries.
- Customizable Matching: Matches resumes based on experience, skills, education, and work history.
- Programming Language: Python
- Libraries:
- LangChain
- Hugging Face
- ChromaDB
- Ollama LLM
- Models:
sentence-transformers/all-MiniLM-L6-v2(Embeddings)deepseek-r1:7b(LLM)
-
Clone the repository:
git clone <repository-url> cd <repository-folder>
-
Create and activate a virtual environment:
python -m venv venv venv\Scripts\activate
-
Install dependencies:
pip install -r requirements.txt
-
Ensure the Ollama API is running locally:
ollama start
Update the CONFIG dictionary in main.py to customize paths, models, and parameters:
pdf_directory: Path to the folder containing PDF resumes.db_name: Name of the vector database.ollama_api: URL for the Ollama API.model: LLM model name.embedding_model: Hugging Face embedding model.
-
Run the application:
python main.py
-
Enter job details when prompted:
- Job title
- Years of experience
- Required skills
- Education requirements
-
The application will analyze resumes and provide the most relevant match.
The system generates a query like:
You are a hiring manager. I will give you a job description and a resume.
For a [Job Title] position requiring [Experience] years of experience,
with required skills: [Skills], and education: [Education],
you will pick the most relevant resume from the list of resumes.
The application outputs the best-matched resume based on the provided job description.
Contributions are welcome! Please fork the repository and submit a pull request.
This project is licensed under the MIT License. See the LICENSE file for details.