My best design work

After spending a year on a product nobody wanted, I did the complete opposite — I picked a market where demand was already proven. Email verification services had millions of paying customers across dozens of competitors. The only way I'd fail was if I couldn't build a quality product and market it.

ListCleanser is an email verification platform. Users upload their email lists, the system runs each address through a 20+ point verification engine — SMTP validation, domain checks, mailbox detection, spam trap identification — and returns cleaned results with deliverability scores.

I didn't write a line of code on this one. I hired an architecture consultant for R&D and proof of concept, brought on a backend engineer for the verification engine, and managed the full build through user stories and structured sprints. My job was the product: every screen, every flow, every pixel. I hired a designer to plan out the first version of the marketing site, and then took it from 80% to 100%.

The design is where I went deep. I designed 99 screens in Figma — the full marketing site, the complete app UI, onboarding, billing, team management, API dashboard, every modal and toast and error state. I'd hired someone on Fiverr for the initial visual direction, then overhauled the design myself. When a freelancer implemented it, I was ruthless about every detail matching the comp.

Something clicked during this project. Visual hierarchy, progressive disclosure, spacing systems, typographic rhythm — principles I'd understood intellectually became intuitive. Figma stopped being a tool I used and became an extension of how I thought about product. I could feel when something was two pixels off, and it bothered me until I fixed it.

To benchmark the verification engine, I set up an eval system that tested our results against NeverBounce, Emailable, and ZeroBounce — the top providers in the space — scoring accuracy across thousands of addresses. We performed very well. That methodology — giving engineers a clear quality target and an automated way to measure against it — has proven to be an instrumental technique that's helped me as I've transitioned into an AI-native SDLC.

The platform is built end-to-end. The verification engine works, the UI is implemented, the infrastructure is robust. But email verification is a low-price-point, high-infrastructure-cost business, and at bootstrapped scale the unit economics were too tight. It's on the shelf for now — the marketing site is still live at listcleanser.com, and the bones are there if I ever pick it back up.