Phygtl: a GenAI social platform

The problem

Generative models made it possible to turn an ordinary photo or clip into something immersive, but nobody had defined what a social product built on that looked like: what creators make, what viewers do, and why either comes back. The company needed a product definition, not just a demo.

What I owned

As the founding design and product lead I owned the product and UX strategy end to end:

• UX architecture. The core object model (what a creation is, how it is shared, remixed, and discovered) and the navigation that followed from it, written down before any screen was drawn.
• Creator tools. Flows that let non-experts produce 3D experiences from their own content, with the generative step framed as a creative choice rather than a black box.
• Onboarding and activation. A first session designed around making something in the first minutes, with the AI doing the heavy lifting and the user making the decisions that felt like theirs.
• Community mechanics. Reactions, remixing, and sharing designed to support acquisition and extensibility, so each creation could bring the next creator in.

Working in ambiguity

Early-stage product work rarely offers clean research or stable requirements. My approach was to keep the architecture explicit and the prototypes real: define the few concepts that had to stay fixed, prototype everything else quickly with the engineering team, and let creator sessions decide. Decisions were written down with their reasons so the team could revisit them when the model capabilities moved, which they did every month.

What I took from it

The founder's seat taught me how much of product design is deciding what the product is under uncertainty, and how valuable it is to be able to prototype the answer yourself. That experience is what I brought to DNNR Tech a few months later.