- Collaborated with the Department of Medical Informatics to analyze stroke prediction data from the MIMIC-IV andTaiwan Health Insurance databases- Applied machine learning techniques including Random Forest, SVM, and XGBoost to develop a stroke prediction model with high accuracy- Addressed data imbalance using hybrid sampling, achieving an AUC of 0.84 and accuracy between 73-77% for different timeframes

- Worked with FSSA’s Deputy Director of Data Science to analyze Indiana 211 caller data and improve health outcome-driven resource allocation- Employed machine learning techniques such as classification to identify patterns in demographic and health data - Utilized advanced models like Decision Trees and SVM to categorize caller needs for targeted interventions

- Supported advertising campaigns, enhancing brand visibility- Built a new version of the company’s database using Python and SQL

- Co-led 30+ year non-profit, managing cross-functional teams in finance, tech, philanthropy, and mar- keting- Orchestrated 2 national conferences, engaging 1300+ members- Initiated 8 online projects, amplifying strategic outreach to 700+ students - Restructured cloud data storage for member accessibility

- Contributed to design and implementation of exam practice websites - Crafted promotional strategies for international summer camp- Enhanced educational subscription services