🤖 AI Resume Matcher using Mistral 7B (Local LLM - No Internet Required)

This project is a Python-based intelligent resume screening tool powered by Mistral 7B Instruct (Q5_K_M GGUF) running locally on your CPU/GPU.
It matches multiple resumes against a single job description, scores them with a match percentage, provides detailed reasoning, and saves the result in Word documents automatically.

---

▶️ Watch Demo on YouTube

---

🚀 Features

• Takes unlimited resumes from the data/resumes folder
• Uses a powerful LLM (Mistral 7B Instruct) locally via ctransformers
• Analyzes each resume with your job description
• Outputs a detailed Word file for each candidate (saved in results/)
• Runs fully offline after setup (no API keys or cloud costs)

---

📂 Project Structure

AI-Resume-Matcher/
├── app.py - Main script
├── reasoner.py - Handles LLM-based reasoning
├── utils.py - Resume text extraction
├── models/ - Place Mistral model file here
│ └── mistral-7b-instruct-v0.1.Q5_K_M.gguf
├── data/
│ └── resumes/ - Add candidate PDF resumes here
├── results/ - Output Word files stored here
├── download_model.bat - One-click model downloader for Windows
├── requirements.txt
├── LICENSE
└── README.md

---

🧠 Model Setup (Required)

GitHub doesn’t allow files larger than 100MB, so you must manually download the LLM model or use the BAT script provided.

Option 1: Manual Download

1. Go to this Hugging Face page:
   https://huggingface.co/TheBloke/Mistral-7B-Instruct-v0.1-GGUF

2. Download the file:
   mistral-7b-instruct-v0.1.Q5_K_M.gguf

3. Place it into the models folder like this:
   models/mistral-7b-instruct-v0.1.Q5_K_M.gguf

Option 2: Use BAT Script (Windows Only)

Instead of downloading manually, run the file:
download_model.bat

It will automatically:

• Create the models folder (if not present)
• Download the correct .gguf model file

---

📥 How to Use

Step 1: Install Requirements
pip install -r requirements.txt

Step 2: Add Resumes
Place any number of .pdf resume files inside:
data/resumes/

Step 3: Run the App
python app.py

You will be prompted to enter:

• The job title
• The job description (paste full text)

Then the script will:

• Parse all resumes
• Match them with the job
• Save a detailed .docx file for each in the results folder

---

📄 Output Example

For each resume, a Word document will be generated with:

• Job title and description
• Resume filename
• Match percentage (0–100%)
• Reasoning explaining:
  • Skills that match
  • Missing qualifications
  • Strengths and red flags

---

📄 Requirements

• Python 3.9–3.11
• RAM: At least 8–16GB
• Works on: Windows, Mac, Linux
• GPU recommended (but not required)

---

🪪 License

This project is licensed under the MIT License.

---

👨‍💻 Author

Made by Aarav Mehta

---

⭐ Star the repo if you find it helpful!