What happens when a PhD researcher builds a product
No VC timeline. No bloat. Building Gaia at the intersection of academic research and real-world use.
Most startups optimize for growth. I'm optimizing for understanding. I'm a PhD researcher at the Technion studying how AI can serve the way architects actually design — not just generate images, but understand spatial logic, material behavior, and the iterative nature of architectural thinking.
The doctorate gave me something most founders don't have: permission to be wrong for longer. You don't have to ship features every sprint because a board wants to see metrics. You can spend three months understanding why a tool isn't working before deciding what to build next.
What research-backed product development looks like
- Countless sessions with practicing architects, students, and firms before building V2
- Every major feature decision grounded in HCI research on design cognition
- Prompt engineering informed by how architects actually describe spatial ideas
- The Knowledge graph emerged from studying how architects manage project context over time
This doesn't mean slower. It means different. When you understand the problem deeply, the features become obvious. You stop building what's easy to build and start building what's hard to build because it's actually needed.
The constraint is a feature
No VC timeline means no pressure to add features that make the deck look good but make the product worse. Gaia has stayed focused on one question: what does an architect actually need from AI during a real design session? Everything else is noise.
The result is a tool that serious architects use for serious work — not a toy you try once. That's the only metric I care about.