AI Product Consulting • Vibe Coding Education
Case files, services, and a few opinions.
Couples visit 5+ venues, spend months imagining setups, and still book blind. No tool let them see their wedding before committing.
Solution: Full-stack SaaS that generates AI venue visualizations — upload a space, describe your vision, and see photorealistic renders of your setup in minutes. Multi-tenant architecture with Stripe billing, role-based access, and a complete admin layer. Built with Row-Level Security in Supabase for tenant isolation, server-side image generation pipelines via Gemini, and transactional email flows through Resend.
✓ Live SaaS product with paying customers
Next.js, Supabase, Stripe, Resend, Apollo, Gemini, Claude
View project →Media planning teams were drowning in manual research — pulling influencer stats by hand, brainstorming in docs, and stitching plans together across a dozen tabs.
✓ Planners got 40% of their week back
Next.js 14, Supabase RLS, Gemini, OpenAI, Recharts, dnd-kit
Learning choreography from videos is hard — you can't slow it down to the beat, and the moves aren't broken into steps. My niece wanted to learn routines for her favorite songs, so I built one.
✓ Built in a weekend, immediately stress-tested by a 9-year-old
Next.js, Spotify Web Playback SDK, LRCLIB, Gemini Vision
Every notes app stores text. None of them help you do anything with it — no connections between ideas, no surfacing what matters, no turning notes into action.
A NYC service business was losing jobs to competitors who responded faster. Quote requests sat in Gmail for hours, pricing lived in spreadsheets, and every response was written from scratch.
People keep hearing "just vibe code it" but nobody actually explains how to set up the tools, save your work, or put something on the internet.
Read more →Figure out where AI belongs — and where it doesn't
1–2 weeks
Most AI features fail because they solve the wrong problem. I look at what your team actually does all day, find the parts where AI would cut real time, and tell you what to build first.
I join your team and we build the thing together
2–4 weeks
I sit with your engineers, we figure out what to build, and I help ship it. When it’s done, I make sure your team knows how everything works so you’re not dependent on me.
I’m the person you Slack when you get stuck
Monthly retainer
Your team can build but keeps getting stuck on AI-specific decisions. I'm the person you Slack when something isn't working.
“If users notice the AI, you've probably over-engineered it.”
Product instinct beats technical demos.
Engineers optimize for what's possible. Product people optimize for what matters. The best AI features come from product instinct, not technical demos.
The real wins are in the boring stuff.
Most companies think AI strategy means 'add a chatbot.' The real wins are in the boring stuff: automating internal workflows, reducing decision latency, and eliminating the copy-paste jobs nobody wants to do.
Build defensibility beyond the model.
If your entire value prop is 'we use AI to do X,' you're one API update away from irrelevance. The moat isn't the AI — it's the workflow, the data, or the distribution.
The tools caught up to the ambition.
Non-technical people can now build production apps with AI. That's not hype — I've watched it happen. The ones who succeed aren't the most technical. They're the ones who understand the problem best.
I'm currently accepting engagements for Q2 2026. Or learn to build it yourself.