7+ years across product analytics, search relevance, fraud analytics, and marketplace operations — plus 10+ years running e-commerce and digital operations first-hand.
Founded and scaled a profitable multichannel vintage apparel business — custom storefront, Depop, eBay, Whatnot, regional markets — generating $60K+ in sales across 1,500+ items, with channel-level reporting guiding sourcing, pricing, and allocation.
Helped scale a Whatnot live-selling channel from ~$30K to $250K+ monthly gross in under 12 months: show structure, curation, bulk listing workflows, and KPI tracing across a ~30-person operation.
Led a commercial AI check-fraud platform from prototype to production handoff — OCR, computer vision, and NLP combined into explainable indicators. Pilots showed ~30% lower review time and 25% fewer false-positive investigations.
Analyzed 2M+ transaction rows with Python, SQL, and Tableau; built fraud-ring dashboards and executive reporting presented directly to the VP and President of Fraud Analytics. Cut manual data prep ~40%.
Improved an internal Apache Solr search platform serving 10,000+ employees — search-log and zero-result analysis contributing to a ~34–35% increase in platform usage.
Photo-first inventory & POS platform for resellers. Next.js/TypeScript PWA — 61 API routes, 16-model Postgres schema, OpenCLIP visual search, Stripe billing. Used live at markets and Whatnot shows.
My own multichannel vintage apparel business: custom storefront, ~2,000-item inventory system, 1M+ monthly Instagram impressions, and services offered to other vendors.
Source for Yanni and Vintage Shirt Tracker — a reverse-image-search pricing tool (FastAPI, OpenCLIP + pHash, Google Lens price aggregation) self-hosted on a Linux VPS.
TypeScript
Next.js
React
Python
FastAPI
Prisma
Supabase
Stripe
Vitest
Sentry
Vercel
Linux / VPS / Nginx
Apache Solr
Open to data science roles in product and marketplace analytics, plus founder and consulting conversations.