I'd rather learn the answer than win the argument.

Social wellness and fitness app (live on the App Store). Promoted from intern to Product Director within a month. Owned the Groups feature end to end — argued for building it on the basis of social differentiation and retention through accountability, then ran vetting, sequencing, build oversight, and testing through to ship. Groups launched, and most users in our testing population used it, with encouraging retention (mixed across cohorts, so I'd call it a strong signal rather than a clean lift). Separately triaged a large, unstructured backlog of user and team requests into priority buckets; shipped fixes included resolving a confusing verification mechanic and cutting home-screen clutter.

A morning dashboard that pulls my Todoist tasks, Google Calendar, Gmail threads that need a reply, and the weather into one screen, with a one-line AI daily summary on top. Next.js and TypeScript, built end to end with Claude Code, Cursor, and Codex across five scoped phases. I built it to scratch my own itch and to stay honest about what AI-native building actually feels like from the inside, since that's increasingly what the teams I'll PM for are doing.

Mentoring two first-time PMs, each leading a different DALI Lab team: an AI learning platform that generates personalized practice for students and gives instructors visibility into progress, and a pair of BRIC games projects. I coach them on requirements, prioritization, user research, and running stakeholder meetings, and I teach PM classes for the lab. Turns out explaining why you'd cut a feature is the fastest way to find out whether you actually know.

A concept I'm testing: a structured app for immigrant parents to learn English with their already-fluent kids as the guides, and immigrant-serving nonprofits as the distribution channel. Right now the work isn't building — it's finding out whether the premise survives contact with real users. Three assumptions are load-bearing and each could be false: that the barrier is a missing tool rather than time or cost or embarrassment, that kids want the teaching role, and that nonprofits will put something concrete behind it. Talking to nonprofits first, then parents and kids. If an assumption breaks, that finding is the point.