-Worked as a graduate researcher on two long-term projects regarding 3D scene understanding-Collaborated with two other lab members to create a joint 3D monocular object detection and scene generation pipeline, implementing a depth-aware 3D feature extraction method for monocular object detection.-Ideated and created a novel 3D multi-object tracking method, using graph attention networks to improve performance and explainability

-Spearheaded individualized customer success and expansion strategies for a voice productivity software startup-Conducted over 20 user interviews and created mockups to gain insights about onboarding flow and desktop app-Created and maintained extensive BI dashboard and analytics, including growth, usage and retention metrics

-One of 75 young professionals selected to study and work in Germany for 11 months-Participating in meetings, language courses, university study and an internship to perform the duty of a civilian diplomat

-Worked with a team of 5 product managers in the Advertising and Emerging Technologies team-Compiled research on discovery+ user behavior and presented to the product organization-Presented proposals on emerging technology innovations to senior leadership

-Product manager of two teams through Bits of Good, Georgia Tech's Hack4Impact chapter (pro bono non-profit software)-Visualized data in Microsoft Power BI for the Boys and Girls Club of Metro Atlanta, saving hundreds of dollars a month-Created an intuitive data story for The Ray, connecting them with important stakeholders-Worked with an engineering manager to create a detailed product feature specification and roadmap for the projects

-Provided data and technology solutions to Democrat Sam Edney's 2020 campaign for the North Carolina House of Representatives-Optimized placement of campaign signs on roads and yards using R and NCDOT traffic data-Predicted early voting turnout by precinct and day using R and historical election data

-Created backend for Suggested Zones V2, an initiative that sends drivers to optimal places and times to drive, increasing driver hours by 21% in experiment groups-Created tables in DynamoDB and corresponding logic to store and retrieve data flexibly-Automated weekly data updates for Suggested Zones V2, greatly increasing efficiency-Collaborated effectively with stakeholders across teams and roles

-Created a preview tool for AMP Story Ads using JavaScript, HTML, CSS, AMP-HTML, and lit-html-Added features such as file upload to existing preview tools-Integrated tool into amp.dev, a developer relations platform, to reach thousands of developers and advertisers-Attended leadership workshops at the Google Women in Engineering (GWE) Summit