Machine Learning-based Text Classification in Recruiting

Automated Talent Matching using Machine Learning

In collaboration with a leading international recruitment agency in Amsterdam's Zuidas district, I co-developed a tailored Data Science solution that delivered over €200,000 in business value within six months.

Central to this solution was a fully automated data scraper (Selenium), combined with an XGBoost-based text classification model. By refining the approach through continuous validation with recruitment experts, we achieved classification accuracy rates ranging from 95% to over 99% (F1-score), exceeding the initial 90% target set by the agency.

Starting as a solo effort, the project later expanded into a collaboration with a fellow Data Scientist. This initiative streamlined the internal talent matching process, significantly improving efficiency and precision in candidate selection.