I work on AI-powered features, experimentation, and the platform systems underneath them. Years shipping product software that has to be both measurable and reliable.
Led Marketplace's first generative AI launches on listing detail pages — Product Insights and Vehicle Insights — reaching 39.4% adoption and 51.9% 14-day retention. Drove 300K+ incremental daily sessions and ~$7M annualized revenue impact across 200+ experiments. Launched Screenshot to Share, scaling to ~1.5M daily uses.
Owned core feature-flagging and rollout infrastructure handling 40M requests per second across a large portion of Dropbox's product surface. Led migration from REST to gRPC. Authored the experiment-expiration system that reduced high-severity incidents by 50%, and a new creation workflow cutting setup time by 30%.
Built full-stack and embedded software for connected-device and Class II medical-device products. Owned deployment pipelines and backend integrations for OTA updates and data workflows across AWS, Django, S3, DynamoDB, Docker, and Yocto.
↳ Full history and shipped work available on request.
I'm a product-minded engineer with 6+ years at Meta, Dropbox, and Promenade. I've shipped AI product launches, run hundreds of experiments, and owned platform infrastructure that handles tens of millions of requests per second. The through-line is the same: move a real metric, reliably, at scale.
I read dashboards before standup. I write the rollout plan before the diff. I treat experimentation as a first-class part of engineering, not a thing data scientists do at the end. The work I'm proudest of is almost always cross-functional — it shipped because someone made the PM, design, data science, and legal conversations easier, not harder.
I care about craft — the tests pass, the latency budget is respected, the on-call runbook actually runs — but I have no patience for craft-as-aesthetic. The bar is whether the thing works, for real users, at the scale it claims to.
Full-stack workout planning and logging app with reusable templates, history tracking, CSV import/export, and SQLite-backed persistence.
Swift iOS app for location tracking, route visualization, and trip export/import. Built for personal use and shared openly.
macOS utility enabling Linux-style window dragging and resizing from anywhere on the window. Open-source collaborator.