CASE STUDIES / MOBILITY MARKETPLACE · MULTI-REGION

REAL-TIMECONFIDENTIAL · ANONYMIZED

One dispatch engine, any city on the map.

A two-sided marketplace matching riders and drivers in real time — engineered so the matching engine, mapping layer, and payouts work identically whether the first city is Austin, Manila, or Nairobi.

2PLATFORMS, ONE CODEBASE
40+COUNTRIES, ONE PAYMENT RAIL
0REBUILDS PER NEW MARKET
1API CONTRACT, EVERY REGION
ride_dispatch.pngSCHEMATIC · DETAILS UNDER NDA

01THE CHALLENGE

Launch anywhere, without re-platforming.

Ride-sharing lives or dies on two things: how fast you can find the nearest available driver, and how naturally the app behaves in a market you haven’t launched yet. A single-region build tends to hard-code assumptions — one mapping provider, one payments corridor, one currency — that become expensive to unwind later.

The brief was a cross-platform iOS and Android app where dispatch, mapping, and driver payouts all work the same way in every market, so entering a new country is an operations decision, not an engineering project.

02OUR APPROACH

Matching as pure geometry.

One React Native codebase ships to both platforms in every market simultaneously. The matching engine treats "nearest driver" as a pure latitude/longitude problem on PostGIS with a Redis geospatial cache — no per-market logic baked in, so it behaves identically everywhere.

An event pipeline decouples trip events from downstream work, meaning a slow regional partner integration can never block dispatch. Payouts run on a single Stripe Connect integration that already operates across 40+ countries, so onboarding a new market’s drivers doesn’t require new payments engineering.

03THE OUTCOME

A launch playbook, not a rebuild.

Stateless services deploy to any AWS region behind the same API contract — a new market launch is a new region deployment, not a new build. Live tracking and in-trip chat ride the same WebSocket layer regardless of local network conditions.

One notification pipeline serves both app stores everywhere, and a single observability pane (Sentry, Datadog) watches every region from one dashboard.

“Nearest driver” is a pure lat/long problem. The moment per-market logic leaks into dispatch, you are building a different app for every country.ENGINEERING BRIEF · DISPATCH CORE

04 — RESULTS

  • Geospatial matching with zero per-market logic
  • Marketplace payouts across 40+ countries on one integration
  • Event-driven pipeline isolating dispatch from partner systems
  • Live tracking & chat over a unified WebSocket layer
  • New-market launch = new region deployment via Terraform
  • Single observability dashboard across all regions

NEXT / CASE STUDIES

Have a system like this on your hands?

BOOK_TECHNICAL_CALL →