I don't just talk about AI. I ship it.

A selection of products I've designed and built end to end, live and in the hands of real users.

SaaS · Hospitality

BarGuard

AI-assisted bar inventory platform that turns counts and POS sales into real variance insight.

Problem: Bars lose money to pour cost and shrinkage they can't see, with manual counts that never reconcile against sales.
What I built: Full SaaS: inventory counting, POS sync, variance analytics, Stripe billing, and an AI assistant. Next.js, Supabase, Vercel.
Outcome: Live at barguard.app serving real venues with automated variance reporting.

Legal Tech

Legal-tech platform helping law firms get found and chosen in the age of AI search.

Problem: Law firms are invisible in AI answers and losing intake to competitors who show up in AI Overviews and chat results.
What I built: Content + GEO engine, AI-visibility tooling, and a marketing site engineered for AI citation. Next.js on Vercel.
Outcome: Live at verdictiq.org with an AI-visibility wedge for legal marketing.

AI Receptionist + CRM

AI receptionist and CRM that answers calls and chats, captures leads, and books appointments 24/7.

Problem: Local businesses miss calls and leads after hours and can't afford full-time front-desk coverage.
What I built: Realtime voice AI (phone) + WebRTC web demo + embeddable chat, lead inbox, pipelines, tasks, and booking. Multi-tenant.
Outcome: Live, handling real calls, web chats, and bookings for businesses.

SaaS · Trading

Behavioral trading journal that turns trade data into discipline and edge.

Problem: Traders repeat the same psychological mistakes because they never see their behavior patterns clearly.
What I built: Journaling SaaS with analytics dashboards, auth, and a content engine. Next.js, Vercel.
Outcome: Live, helping traders review and improve execution behavior.

AI · Gaming

AI coaching that analyzes gameplay VODs and returns timestamped, actionable feedback.

Problem: Gamers plateau because reviewing their own footage for mistakes is slow, vague, and hard to act on.
What I built: Twitch VOD pipeline + Gemini analysis producing timestamped coaching. Clerk, Neon, Drizzle, Next.js.
Outcome: Live at gamesenseai.pro turning VODs into coaching breakdowns.