Marketplace PM (TikTok Shop, Walmart), Tuck MBA. Everything here works. Some of it works well.
| Project | What it is | Stack | Status |
|---|---|---|---|
| Launchgate | CI for marketing journeys. Lints a lifecycle journey for broken logic, consent and frequency issues before it reaches customers, applies fixes, then launches it to a cohort | Next.js 16 · React 19 · TypeScript · Tailwind | Live · demo |
| VisitBrief | A patient-community-signal prototype that explains why each flag fired and drafts a visit brief (fictional data) | HTML · JS | Prototype |
| Fortune Ledger | An anti-impulse-shopping Chrome extension that adds deliberate friction before you buy | Chrome MV3 · TypeScript | Prototype |
| Weekly Health Brief | Wearable exports → a daily model → a coach-style weekly brief (synthetic data) | Python · matplotlib | v1 |
- Prototype first. A clickable build in days beats a spec in weeks.
- Evals over vibes. Define what "good" means, measure it, and only then scale it.
- Humans in the loop where it matters. Approval gates, consent checks and audit logs are built in from the start.
- Explainability as a feature. Every AI output shows why it happened.
- Honest numbers. If a metric isn't measured yet, it says TBD.
- Agentic Commerce Observatory: tracking how AI agents shop, compare and check out across marketplaces.
- Marketplace agent experiments: buyer and seller agents, trust and friction design, and what changes when the customer is an agent.
Open to AI product roles in commerce & marketplaces · Seattle / remote Personal projects; not affiliated with or endorsed by any company.