At Netflix, I replaced legacy payments models with a simpler, flexible set covering transactions worldwide, from signup and renewals to refunds, chargebacks, and disputes. That cut the time to ship a model change from 1–2 weeks to about a day.
I also built production payments tables and led the migration off the legacy stack, retiring about 65K lines of pipeline code. To improve reliability across engineering pods, I authored data development standards and built a Python testing framework used as a CI gate.
Previously, at Etsy, I built an event-driven pipeline that sent conversion data to Google Ads and drove several million dollars in incremental gross merchandise sales. My work has also spanned aviation and energy data science, including maintenance insights and monitoring algorithms; more recently, I’ve applied MCP servers and multi-agent workflows with Claude and Codex to help teams ship faster.
