EURES Job Scraper, AI Matcher & Web Viewer

This Python-based system automates scraping job listings from the EURES portal, stores them in SQLite, matches them against a candidate’s resume using advanced AI models (OpenAI, Google Gemini, LLaMA), and provides a Flask web app to explore and filter matched jobs.

---

🚀 Features

✅ Job Scraping & Storage

• Scrapes job listings from the EURES portal.
• Manages cookies and XSRF tokens automatically.
• Handles pagination to fetch all jobs.
• Saves structured job data to a SQLite database.

📄 Resume Matching

• Extracts text from resume PDFs.
• Matches resume to jobs using a semantic embedding model.
• Produces top N matches.

🤖 AI Reranking

Rerank top job matches using LLMs:

• LLaMA CPP (local): Efficient binary prompt reranking.
• Google Gemini 2.5 Flash: Fast and detailed job-to-resume evaluation.

Each AI model provides:

• Match score (1–10)
• Justification for score
• Extracted contact person/email
• Draft email for application

🌐 Flask Web App

• Paginated UI: Browse all jobs and matched jobs.
• Job Detail Page: View full job details with match scores and AI-generated metadata.
• Filtering: Filter jobs by location and other fields.
• Sorted View: Display top matched jobs sorted by reranked score.

---

🗂 Directory Structure

├── main.py                   # EURES job scraper entrypoint
├── rank_llama_cpp.py         # AI matching & reranking logic
├── app.py                    # Flask web app
├── templates/
│   ├── index.html            # Job list view
│   └── detail.html           # Single job view
├── static/                   # CSS/JS assets (if any)
├── config.ini                # Paths and settings
├── jobs_data.db              # SQLite database
├── resume.pdf                # Resume file for matching

---

⚙ Configuration

config.ini example:

[Paths]
DB_PATH=jobs_data.db
RESUME_PDF_PATH=resume.pdf
LLAMA_ENDPOINT=http://localhost:8090

[Settings]
TOP_N_MATCHES=10
FINAL_TOP_N=3
N_PREDICT=32

[Models]
SBERT_MODEL_NAME=all-MiniLM-L6-v2
LLAMA_MODEL_PATH=path/to/llama/model
; OPENAI_MODEL=gemini-2.5-flash
OPENAI_MODEL=gemma-3-27b-it

[API]
OPENAI_API_KEY=your_openai_api_key

[EXTRA]
NAME= Your Name
TARGET= Desired Job Title

---

💻 Installation

1. Install Python dependencies:

   pip install -r requirements.txt

2. Install ChromeDriver and place it in your PATH.

3. (Optional) Install uv if using it for running:

   pip install uv

---

🧪 Usage

Scrape Jobs

uv run main.py

Match and Rerank (Example)

uv run rank_llama_cpp.py

Launch Web App

uv run app.py

Then open http://127.0.0.1:5000 in your browser.

---

🔍 Web App Features

• /: View all scraped jobs (paginated).

• /job/<id>: View full job details, including:

  • Description
  • AI score
  • Justification
  • Contact person/email
  • Email draft

---

🔧 Troubleshooting

• 403 Errors: Clear cookiefile.json and xsrf_token.txt and retry.
• Selenium issues: Ensure ChromeDriver matches your Chrome version.
• LLaMA model path errors: Double-check model and server paths for local models.

---

📜 License

MIT License. Use, fork, extend freely.

---