Developed backend and machine learning (ML) for a full stack, multi-layered, POC web application that expedites risk analysis of Capital One's car dealership partners by aggregating, filtering, and sorting 12 million google reviews for 16,000+ dealerships from the Google Places API; layers developed include micro-frontend UI, BFF, batch job, database, search API, and classifier. Utilized JPA repositories and Java Database Connectivity connections to sort and filter all dealership reviews by dealer name, rating, review category, time submitted, and/or predicted risk level under 750ms. Designed a supervised ML logistic regression model to predict potential risk of dealer misrepresentation for 12 million reviews, reducing false positive by 75%. Presented 8 technical demos to over 100 stakeholders including VPs and was immediately approved for production.

Used SQL database management and .NET Framework (C#) to design websites and desktop applications; connected SQL databases to visual C# through both model browsers and Open Database Connectivity connections. Developed a Query By Example Windows application using C# to allow users to create, edit, delete, and execute saved queries pertaining to any stored procedures, views, or tables in the local SQL server while choosing which parameters to include. Created an ASP.NET web form to allow users to create, edit, and delete orders and order lines; a returning customer’s order willupdate the customer’s account in the SQL database while a new customer will have an account set up for him/her

Created an administrative website for a financial firm to systematically store client information in an organized manner, increase efficiency, and obtain backups of all data. Modernized the company’s record-keeping system by digitalizing their hard-copy files. Taught other staff members to properly issue checks and record deposits using QuickBooks to manage company expenses andmonitor cash flow