EatIn App

A two-sided marketplace connecting home cooks with nearby eaters — supporting dine-in, pickup, and delivery. I owned the full stack end-to-end: product architecture, design system, mobile and web frontends, backend services, and CI/CD pipeline, under a 4-month delivery target.

Full-Stack Engineer & UI Designer

../Responsabilities..

Clients

../Stack...

React Native / React / TypeScript / FeatherJS / Socket.io / JWT Auth / Redux / Styled Components / MongoDB / Google Cloud Platform / GitHub Actions / Figma

Insights

EatIn was the most concentrated exercise in end-to-end ownership I’ve had. When you’re the only engineer making architecture decisions, there is nowhere to defer: every technology choice, infrastructure trade-off, and UX decision lands on you.

The constraint — a functional prototype in under four months — forced a discipline that large-team projects often lack: ruthless prioritization. I mapped the critical path on day one and worked backward from it. Features that didn’t serve the core user flow were deferred without negotiation. Tools were chosen for what they could deliver today, not for what they might scale to hypothetically.

The most technically demanding piece was the real-time layer. A two-sided marketplace means two independent sessions — cook and eater — need to see the same state without latency. I implemented WebSocket event handling with explicit acknowledgement patterns to prevent silent failures, and tested concurrent update scenarios before considering the feature complete.

The design system was a forcing function for the frontend architecture. Building it in Figma first — with tokens for color, spacing, and type — meant the component library had a contract to implement against, not just a mockup to approximate. That discipline is one I’ve carried into every project since: the interface between design and engineering should be a specification, not an interpretation.