A movie recommendation engine built around your Letterboxd history. Syncs your ratings and watchlist, builds a taste profile from genre affinity, keyword clusters, and critic alignment, then ranks your watchlist by predicted enjoyment. Built end-to-end — data pipeline, ML inference, and web app.
Selected work from Intuit Mailchimp. Code is proprietary — shown here as recorded demos.
Built mailchimp.com's first personalization system for prospective visitors. Captures preferences on first visit and immediately updates homepage content client-side — loading only what each visitor actually sees. Still running in production.
Led engineering on a 14-module redesign of mailchimp.com's homepage. Coordinated across marketing, product, and design to phase overlapping A/B tests so the data stayed clean without slowing the launch timeline.
Led a 3-engineer team building a React module that walks prospective customers through creating a live email campaign in real time. Shipped to mailchimp.com and drove $280K+ in incremental annual booking revenue.