Diana Xu

- Spearheaded a team of 3 to develop a full-stack serverless medication tracking application from 0 to 1 for a healthcare technology startup venture HugMed backed by Cornell University, amplifying medical adherence for elder patients and caregivers - Implemented and updated user authentication, medication management and tracking features with Agile methodology using Vite, React, Node.js, reducing manual intervention by 50% - Leveraged AWS (including API Gateway, Lambda, and DynamoDB) and Supabase services to create a highly scalable and reliable backend infrastructure, enabling seamless user interactions and real-time medication tracking - Redesigned database architecture and optimized SQL schema, achieving a 10% reduction in overall query time - Collaborated with cross-functional teams by using Trello; Initiated UI wireframes in Figma to support product pitches, directly contributing to successful securing of $5,000 in entrepreneurship awards - Created clear source code documentation and user journey maps, refining patient-caregiver workflows for improved usability

- Optimized machine learning models to predict book award winners under the consideration of Time Series with Dask, Numpy, Pandas, and scikit-learn, achieving a 31% match rate between predicted and actual top 100 books using Random Forest Regression - Designed and implemented an ETL data pipeline, migrating data from multiple databases, performing data cleaning, and transforming data to feed into machine learning models - Identified correlation among 60+ continuous, ordinal, and nominal features and produced data visualization to support feature selection, data extraction, and data cleaning with 10+ million original real-world data by Python - Presented strong professional communication skills by conducting weekly meetings and team cooperation

- Designed a web application tool in Flask for data analysis and visualization, enabling XML data uploads and automating data processing by integrating Python and R environments, reducing manual effort by 40% - Optimized workflow to extract and synthesize metrics from 400+ research data files simultaneously within 8 seconds - Performed clear code review of in-house medical data search engine on GitHub, reorganizing code structure and enhancing maintainability by 15%

- Conducted Network Meta-Analysis(NMA) based on clinical data to help compare multiple drugs across different endpoints - Automated data quality checks and streamlined reporting pipelines, improving data processing speed by 25% while ensuring data integrity and reproducibility for regulatory review