• Led product analytics for a new server-to-server 3rd-party ad platform at eBay, designed A/B tests and drew insights that helped the product scale and improve its performance• Built ETL pipelines and dashboards to track the key performance metrics of the product• Explained an ad request gap by analyzing event log and attributed it to users clicking the back button when navigating between pages on eBay, which increased annual ad revenue by ~$4M• Identified a problem that prevented ads from rendering and found the root cause via deep dive analyses, which increased ad revenue in 2018 by ~$5M• Analyzed and attributed reasons underlying sellers churning from our 1st-party ad program• Designed a framework to detect spikes of given search queries/keywords using time series analysis, fully put it into production using Python/Scala and built a dashboard for visualization• Managed 10+ stakeholders from teams globally

• Built MusiCon, a web app to automatically detect and extract music clips out of a YouTube video• Developed a machine learning model using principal components analysis and logistic regressionthat achieved an AUC of 0.90 on test data of music and non-music audios from YouTube• Engineered the front end using Flask and JavaScript, and deployed the app on AWS EC2

• Led data analysis of our research group for a multi-center genome-wide association study (the GUARDIAN consortium) on type 2 diabetes and analyzed large datasets with a size of ~10GB• Programmed a pipeline for genome-wide association analyses using Bash/Perl/R, and parallelized the computing using pbsdsh, which increased efficiency by >8 folds• Participated in weekly conference calls with collaborators on behalf of our research group• Co-first authored a paper based on the results, with several other papers in preparation• Had poster presentations on behalf of our research group at the annual conferences of ADA and ASHG.