CrewLink
The problem
Airline crew in Dubai work rosters that no consumer ride-hailing product is built for: 4am pickups, layover-shaped demand, and the same handful of corporate sites over and over. Nobody had built a dedicated platform for it.
What I built
The whole system, from first commit: rider apps, driver and dispatch tooling, the admin dashboard, subscription billing, loyalty and reporting — plus a proprietary matching engine that decides who shares a ride, which route is taken and what each seat is worth. Kept in-house as trade-secret IP rather than licensed in, because the platform's retention economics sit downstream of it.
Scale
Tens of thousands of completed trips across dozens of live service areas in the UAE.
Retention
A repeat-purchase rate well into the eighties — the platform is a weekly habit, not a one-off.
Growth
Order-of-magnitude ride growth inside the first year of operation.
Reliability
A trip-completion rate above ninety percent on time-critical airport runs.
Also shipped
- A subscription and loyalty engine — recurring plans, ride allowances and earned credit — that carries a large share of active riders.
- Shared-ride economics riders actually opt into, with pooling logic that balances match rate against detour time.
- Demand modelled as a hub-and-spoke network anchored on major airline corporate sites, extending into neighbouring emirates.
- Production robustness controls across PCI-grade payment handling, bank-grade identity verification and fraud prevention.
- Substantial quarter-on-quarter organic search growth on zero paid spend, through technical SEO on a Next.js stack.