Ezequiel Perez

At Hackaton ONE x NoCountry, I participated in the collaborative development of a churn prediction solution. I worked on exploratory data analysis, feature engineering, and classification models, applying cross-validation and hyperparameter optimization with a focus on recall.

I developed an end-to-end anemia prediction system using hematological parameters. Its CatBoost pipeline achieved 0.99 recall on the test set, with SHAP explainability; I also implemented a REST API and prediction persistence.

For Telecom X, I developed a churn modeling project that compared classification models and documented the methodology, results, and conclusions. My work across these projects includes Python-based data analysis, machine learning, and model evaluation.