A platform for
working artists
A rebuild and a rebrand at the same time, an engagement that had just lost its design lead, and thousands of artists who all want their work shown differently.
A small company runs the websites of thousands of working artists. It was rebuilding its application and rebranding at the same time, and the engagement had just lost its design lead. Artists are particular about how their work is shown, for good reason, and half of them wanted the opposite of what the other half wanted. The editor had grown options faster than anyone could find them.
I owned the product vision for the parts artists touch most, the editor and the page that shows a single artwork. The rules for what an artist can configure and where that choice lives. And from August 2026, the design direction of the marketing site for the new brand.
Don't design what
the data can't hold.
Every prototype stayed anchored to what the platform actually stores. If a design needed a field that didn't exist, we labeled it as proposed instead of quietly drawing it. I also had the prototypes built from the production code's own components, so engineering got something they could paste rather than reinterpret.
Steady it before changing it.
I came in as the engagement lost its design lead. The first job was trust: research with artists, a playback the client could act on, and a clear vision for the areas the rebuild would touch first. My former head of design described it in a review as helping "stabilize the engagement" and "creating a new vision for key areas of the product."
One good default, then controls.
Any single layout is a compromise when the artists split down the middle. So we split the work: phase one ships one strong default for everyone, phase two gives artists control.
For the controls, I chose a site-wide default with an optional per-gallery override, not a setting on every artwork. Layout is appearance, it changes rarely, and consistency is worth a little friction. The features underneath became independent switches that inherit unless an artist sets them.
Take it when it needs a lead.
When the contract designer moved off the marketing site in August 2026, I took design direction of it for the new brand and brought it to a structure and pricing the founder approved, then handed it to the dev team with written documentation.
How the team grew.
Early on, an experience designer ran the first concepts. I came in to set direction for the whole product, and we brought on a product designer who owned most of the look and feel, working with me and the product manager. My job was to keep the direction steady so they could go deep.
Built on production
Prototypes built from the platform's own code, so engineering could take the work straight into the product.
One settings model
A site-wide default, per-gallery overrides, and switches that inherit until an artist changes them.
Approved and handed off
The marketing site's structure and pricing approved by the founder, documented for the dev team.
Check production first.
I'd check production before the first sketch, not after. An early editor concept put the settings on the wrong side of the screen, because it was drawn from memory of the product instead of from the product. From then on every round started from a fresh pull of the production code. It cost an hour a round and saved every review after it.