Hey, I'm
Hayatu.

I spent three years making things look right before I ever made them work. Two years in graphic design, one in product design, and then I crossed over into engineering and never left. Most engineers learn the user last, if at all. I learned the user first, and seven years later it's still the order I think in: the person on the other side of the screen, then the architecture underneath them.

That order has held up across some very different rooms. As founding engineer at X-HR, it meant taking an AI-native HR product from nothing to MVP and building the frontend platform and internal SDK it ran on. At Tamara, it means checkout and merchant onboarding for fintech that more than ten million people use, where the interesting problems are the ones nobody notices because the system just holds. In between: booking, humanitarian aid, stablecoins. The domains kept changing. The job didn't. Walk into ambiguity, make the calls, ship something that works, stay until it works quietly.

Since then I've helped found teams, made hard calls under real ambiguity, and shipped things that had to hold up at scale. Fintech used by more than ten million people at Tamara. An AI-native HR product I took from zero to MVP as the founding engineer at X-HR. Booking, humanitarian aid, and stablecoin platforms in between. Some of it was genuinely stressful. Most of it was worth it.

Lately the layer I care most about is the one between the user and the model. I've built tool-grounded agents, multi-pass LLM review for pull requests, an assistant living inside the HR product itself. It's the same problem I've always worked on, wearing new hardware: something powerful on one side, a person on the other, and the whole job is making the two make sense to each other.