- Engineered machine learning model to predict sepsis in 75,000 patients from demographic data, vital signs, and lab features using logistic regression, random forest, LASSO, XGBoost, CatBoost algorithms, achieving an accuracy of 0.93.- Performed comprehensive data preprocessing, including exploratory data analysis, balancing, cleaning, and imputation to ensure high-quality input for the model, significantly improving reliability and accuracy of predictions.- Communicated methodologies and insights through biweekly presentations to coworkers, mentors, and fellow interns.