Built and optimized algorithms that serve ads and product recommendations to millions of eBay users across homepage, feed, checkout, and item pages, directly driving conversion and GMV at scale. Led system architecture and end-to-end implementation of eBay's shoppable Live Video carousel, bringing real-time video commerce to product pages and driving ~2.7M annualized views plus ~$4.7M GMV.

Built an AI-native academic co-pilot from 0→1, leading a 5-person engineering team through design, architecture, implementation, enterprise sales, and fundraising. Closed $48K contract and multiple $7K paid pilots with institutional clients by delivering tailored solutions for academic support resource centers, including custom retrieval pipelines and multimodal AI learning workflows.

Built internal debug tooling for Snap Map's OpenGL layer (the rendering engine powering bitmoji, places, shops, and actionmoji), improving visibility into rendering pipeline issues for the engineering team. Delivered full-stack features on Snap Maps (TypeScript, Python, Java, Kotlin) reaching 300M+ daily active users.

Drove test planning and spec review across multiple product areas, including Snap Maps, Profile, and Commerce, accelerating release confidence for Snap's highest-traffic consumer surfaces. Developed entire test suites for new features on Android (Espresso) and iOS (XCUITest) in Objective C, Java, and Kotlin, reducing manual QA overhead and minimizing regression risk.