AI tools like Claude Code, Cursor and Lovable build interfaces fast, but the results often look “almost right”: inconsistent spacing, weak hierarchy, off-brand components, missing states and accessibility failures. I fix that.

I’m a senior product designer with 8+ years of experience. Most recently I was the sole designer at an AI startup, where I reviewed AI-generated UI every day, shipped 20+ pull requests a month, and built an automated audit that catches design system drift and WCAG contrast failures.

What you get • Full audit of up to 15 screens (desktop or mobile) • A prioritised issues report covering layout, hierarchy, spacing, typography, interaction states, accessibility (WCAG AA/AAA) and component consistency • Redesigned Figma screens for your 5 most important screens • A clean design system foundation: colour, type and spacing tokens plus core components • A recorded Loom walkthrough of every fix and why it matters

Optional add-on: I can implement the fixes directly in your codebase with Claude Code and open reviewed pull requests on GitHub.

Ideal for: founders and teams who built an MVP quickly with AI and now need it to look trustworthy before launch, fundraising or onboarding customers.