Amy Weitzman

Bachelor's

Computer Science

Behaivior

Software Developer and Data Science Intern

• Streamlined daily survey process by designing and implementing custom survey component using Figma, TypeScript, React.js, Redux, Firebase yielding 3x better response rate and improving data collection efficiency by 50% • Composed event-driven Firebase Cloud Function to automatically send email alerts to care provider when person-in-recovery indicates need for assistance, saving company 10 man-hours per week • Visualized real-time user health information collected from wearables using Victory charts library in React.js and Firebase to monitor potential onset of craving in 1 and 6 hour intervals • Cleaned dataset of 1.2 million time-series data points to standardize format and aggregate 8 features by time intervals • Trained gradient boosted decision trees (XGBoost) classification model in Python to predict whether person-in-recovery will be craving in the next 6 hours with >90% accuracy