At Amazon, I established an operating model for auditable AI agents and scaled it to govern four production agents globally. I also architected and built a profitability decision agent that evaluates more than 25,000 at-risk catalog items monthly.
I built an automated evaluation harness that benchmarked about 800 outputs against a golden dataset, achieving 86% accuracy. I also shipped a full-stack LLM agent for Business Reviews, eliminating more than 320 hours of recurring manual preparation.
Earlier, I worked on shopper and seller products at Samsung, Wayfair, and eBay, including a Galaxy Watch configurator that lifted add-to-cart engagement 63%. In my personal projects, I’ve shipped AI products including GreenLens, a landscaping estimator live with pilot users, and Uplift, a voice-to-voice AI companion.

