AIJobMate

AIJobMate is a prototype application for generating UK-style CVs and cover letters using structured personal data and multiple AI agents orchestrated via CrewAI.

🔧 Overview

AIJobMate works in two main steps:

1. Build Profile
   Enter your professional background in freeform text. The system extracts structured data (experience, skills, education, etc.) and stores it in data/profile.json.

2. Generate CV & Cover Letter
   Provide a job description and select which model to use for each agent:

   • 🧑‍💼 CV Writer
   • 📝 Cover Letter Specialist
   • ✅ Quality Assurance Reviewer

Each agent uses a selected model (e.g. llama3.2, llama3.1) and contributes to creating well-structured documents.

✨ Features

• Structured profile generation from natural language input
• CrewAI-powered workflow with distinct agent roles
• Independent model selection per agent
• Quality review step before final output
• Ready to run locally with Ollama
• Modular codebase for future expansion

🚀 Getting Started

1. Clone the repository

git clone https://github.com/loglux/AIJobMate.git
cd AIJobMate

2. Set up your environment

python -m venv .venv
source .venv/bin/activate  # On Windows: .venv\Scripts\activate
pip install -r requirements.txt

3. Start your local Ollama server

ollama serve

Pull models you want to use, for example:

ollama pull llama3.2

4. Run the application

python gui.py

🖥️ Interface

The Gradio interface has two tabs:

🧱 Build Profile

Input: Your background in plain English — this can be copied from your CV, or written freely with comments and additional details. Output: JSON profile stored in data/profile.json
Example:

I’ve worked in IT operations for several years, including network monitoring, scripting with Python, and automation using Docker. My most recent role involved collaboration with cross-functional teams. I’d like to highlight my ability to solve complex technical problems and communicate with both technical and non-technical stakeholders.

P.S. 🛠️ The Profile Builder currently uses the llama3.2 model by default to extract structured JSON data from free-form text input.

📄 Generate CV & Cover Letter

Input:

• A job description
• Selected models for each agent

Output:

• CV
• Tailored cover letter
• QA-reviewed summary

📁 Project Structure

.
├── gui.py                  # Gradio interface
├── career_crew.py          # CrewAI agent and task logic
├── profile_manager.py      # Handles reading/writing profile JSON
├── llm_engines/
│   └── ollama_client.py    # Low-level client for Ollama models
├── data/
│   └── profile.json        # Generated user profile
├── requirements.txt
└── README.md

⚠️ Note

This is a functioning prototype and will continue to evolve. Expect future enhancements including:

• Editing specific sections of the profile
• Adding custom categories (e.g. Projects, Publications)
• Exporting to PDF or Word
• Support for additional LLM providers (OpenAI, DeepSeek, Mistral, etc.)

📌 Model Notes

You can use any local model available in your Ollama setup.
Examples:

• llama3.2
• llama3.1
• Any other model pulled via ollama pull

No specific model recommendation is made.

📜 License

MIT