Hedi Ghorbel

At Colombus Capital, I advise FX clients on competitive exchange rates and hedging strategies while producing performance reports and market-analysis dashboards. I also contribute to an FX pricing tool that estimates the Central Bank of Tunisia fixing through data analysis, statistical modeling, and forecasting.

For my final-year project at Sofrecom Tunisie, Orange, I processed more than 35 million login attempts with PySpark and built a semi-supervised fraud-detection approach. Using pseudo-labeling and machine learning, the work identified 11% of safe connections as fraudulent attempts.

I've also analyzed more than 2 million customer comments at Sofrecom using RoBERTa and zero-shot classification to identify telecommunications issues, then developed Tableau dashboards to understand negative-feedback causes.

My engineering training in financial modeling and actuarial science is grounded in mathematics, programming, credit-risk modeling, portfolio optimization, and machine learning research. I'm especially interested in applying data science to finance and insurance.