My master’s thesis at Universidad Complutense de Madrid focused on a cardiovascular disease risk and prediction model for early detection and risk classification using an imbalanced medical dataset.
I prepared clinical data with winsorization and log transformations, then selected features using Stepwise AIC/BIC and MMPC. I evaluated XGBoost, CatBoost, Random Forest, Neural Networks, and Logistic Regression with repeated cross-validation and probability threshold tuning to minimize false negatives.
As a freelance data analyst in healthcare, I cleaned and unified operational records into a central database and developed metrics on consultation times and operational efficiency for visual reports.
I also built a demand forecasting model for Seoul Bike Sharing that reduced test error by 26%, and developed a recommender system using vector search and RAG. I’m available immediately and require UK Skilled Worker visa sponsorship.
