AI-Powered Ride-Booking Platform
Cabsmart
Backend systems for an Australian ride-booking SaaS platform — booking, dispatch, pricing, and real-time trip tracking at scale.
role
Senior Backend Engineer
period
Aug 2025 — Present
client
Cabsmart Mobility Pty Ltd. (Australia, Remote)
daily rides — target scale
25K
LaravelPythonVue.jsReact.jsAWSRedisSocket.ioMySQL
the challenge
The platform needed to automate ride booking end-to-end — including phone-based bookings — while keeping dispatch and pricing accurate in real time, and do it on an architecture that could grow from ~1,000 daily rides to 25,000 without a rewrite.
the approach
- Designed and shipped 100+ REST APIs covering booking, dispatch, pricing, payments, notifications, reporting, and trip management, consumed by both the React.js customer app and internal dispatch tooling.
- Built an AI-powered Auto Booking Engine and IVR Booking System using Python-based AI services wired into telephony, so customers could book a ride by phone with no human agent involved.
- Added Call Masking to keep customer and driver phone numbers private during the booking and trip flow.
- Moved real-time state — driver locations, live trip status, dispatch events — onto Redis and pushed updates over WebSockets/Socket.io instead of polling, with Queue Jobs handling everything that didn't need to block a response.
results
- Backend performance improved 30–40% through database optimization, Redis caching, indexing, and API refactoring.
- Architecture validated to scale from 1,000–1,500 daily rides to a 20K–25K target without structural changes.
- Secure, real-time authentication and ride tracking shipped in the customer-facing React.js app.