Most "AI MVPs" stop at Figma flats and dotted-line user flows.

I built a real one — design system, working frontend, live AI pipeline, published output — in days.

Kryscapes is a travel advisor's co-pilot. She records a voice note about a trip. The system writes a polished editorial guide — hotel block, affiliate items, SEO meta, social captions. She edits in place. Hits publish.

What's inside the case study: → PRD + DESIGN.md that scoped the build → Brand → Alias → Map token cascade in code → Three Supabase tables — narrow on purpose → AI pipeline with Zod-validated structured outputs and a deterministic fallback → Processing screen, draft review, published guide proof

You don't leave the sprint with a deck. You leave with a working artifact, the codebase, and a story you can use for pitching or fundraising.

Two ways to hire me: — AI Product Design Sprint · $2,400 — strategy, UX, Figma, prototype, AI workflow map — AI MVP Sprint · $6,500 — design sprint plus working MVP, database, AI pipeline, launch-ready handoff

Read the full case study: Check it out

top class work 🔝

Built an interactive Streamlit dashboard that models revenue recovery scenarios for e-commerce stores. Allows filtering by customer state, product category, and payment type. Three interactive sliders model recovery assumptions — late delivery reduction, customer churn rate, and cancellation recovery rate — and instantly recalculate dollar outcomes. Identified $333,874 in structural revenue risk across 99,441 orders with recovery potential of $40K–$137K depending on execution.

great job 👌