I joined as the founding product designer and owned design well beyond the individual contributor scope. I built the design function from the ground up (strategy, team, system, and process) while simultaneously leading the full platform redesign. This was a dual mandate: ship a better product and build the infrastructure for the team to keep shipping at quality.
Layerr had traction but no design foundation. When I took ownership, the product had unclear direction, a broken UI, and a UX that exposed blockchain complexity instead of abstracting it. I defined what needed to be true structurally before any surface-level design work could hold.
The overhaul touched every layer: information architecture, interaction model, visual system, and component infrastructure. The standard was simple. The platform had to feel Web2-native to users while running Web3 underneath. Every design decision was evaluated against that constraint.
I architected Layerr's design system from zero: token system, component library, style guide, and contribution governance. The system became the team's shared contract, so consistency held and the same problems stopped being re-solved as the platform grew.
The interactive prototype used to pressure-test project and collection management with product and engineering before build. This is a walkthrough, not a controlled experiment. The platform had no traffic split at this stage, so the flow was validated against partner feedback and internal review rather than measured.
Light mode was a product decision, not a design exercise. The dark-only interface was a barrier for mainstream adoption. I built both themes as a token-based system: switchable modes off one component architecture, so any new component supports both automatically. Any new component supports both themes automatically.
Layerr had traction and no foundation. The real question was not what the new interface should look like. It was whether to fix the structure underneath it now or later.
I restructured the information architecture and the component layer at the same time as the visual language, so the new look and the new structure shipped as one thing. The interaction model was rebuilt around what a mainstream user expects, with the blockchain mechanics held underneath it.
The faster route was a visual refresh on the existing structure. It would have shown progress in weeks, kept engineering shipping the whole time, and let the IA work happen after launch, which is where that work usually goes to die.
A long stretch with nothing releasable. The redesign could not go out incrementally, so engineering held while the component layer was rebuilt and the team spent that time without the morale of shipping. It also meant re-cutting flows engineering had already built and tested, which is an expensive thing to ask of people who did nothing wrong. I took that trade because a reskin over the old IA would have locked in the structural problem for another year.
I no longer have access to Layerr's dashboards, so this is what I can still stand behind: one tracked figure, and a set of outcomes with witnesses.
Engagement was tracked and moved; almost nothing else was. I have no figure for activation through the create-collection flow, drop-off at wallet connection, support volume tagged to onboarding, or the share of shipped components coming from the library rather than one-offs. That last one is the honest measure of whether a design system is working, and it is the one teams skip.
I built the design function and led the platform redesign at the same time, which meant the governance model was written by the person with the least time to enforce it. Contribution rules that live in one person's review queue are not a system, they are a bottleneck with documentation. It was also my first time leading, and I under-delegated early. I kept the hardest surfaces for myself when handing one of them over would have grown the team faster than my version of it being slightly better.
Impact: Layerr went from a product with no design foundation to one with a real system, a structured team, and a platform architecture capable of supporting mainstream scale. The design system took repeated work out of the sprint cycle. The interaction model became the template for onboarding new feature surfaces.