Map-First Real Estate Marketplace — Architecture to Production

Built a complete property marketplace from scratch: geospatial search, floating-pin discovery, landlord workflows, subscription-based access, and mobile-first UX shipped to production.

Led end-to-end technical delivery: architecture, backend APIs, geospatial search, interactive map, landlord and seeker workflows, subscription monetization, payment integration, and production deployment.

A property seeker opens the app, drops a pin on any street, and properties appear around that point; filterable by price, category, and availability. They view verified listings, see landlord response rates, and request a callback without touching a broker. On the landlord side: publish a property, set a price, choose a plan. Contact reveals, tour scheduling, and payment access are handled by the platform.

Google Maps was ruled out on cost. Mapbox was the natural alternative, but it has search coverage across African regions, which is exactly where the product was launching. OpenStreetMap has strong local coverage and no licensing cost, but its default UI is too rough for a consumer product.

We built a hybrid: Mapbox GL for rendering and visual experience, OSM for geocoding and search. Users get a polished map with reliable local search, at a fraction of the cost. This required testing coverage across target regions, evaluating three providers, and architecting the integration so both layers stay in sync across pin-drop, search, and filter interactions.

Stack: Next.js, React, TypeScript, Node.js, PostgreSQL (PostGIS), Mapbox, OSM, Redis, Stripe, GCP, Docker, Nginx