CabPool — Dual-Role Ride Sharing with Route Optimization

Transportation CabPool (Internal Product) • 4 months • 6 engineers

The Challenge

Unlike traditional ride-hailing apps with separate driver and passenger apps, CabPool allows any user to switch between driving and riding within the same application. This dual-role model requires real-time matching algorithms that consider both route overlap and schedule compatibility, along with a fare calculation engine that handles pooled ride splitting.

Our Approach

We built the app in Flutter with Google Maps SDK for route rendering, geocoding, and ETA calculation. The matching algorithm runs on the Node.js backend, comparing ride requests against available drivers using geospatial queries (MongoDB 2dsphere indexes) and route overlap scoring. Fare calculation accounts for distance, time, vehicle type, and pool occupancy with transparent breakdowns shown to all participants. Firebase Cloud Messaging handles real-time notifications for ride matches, and in-app chat uses Firebase Realtime Database for driver-passenger communication.

Tech Stack

Flutter Node.js MongoDB Google Maps API Firebase

Key Results

Dual-role switching with zero app reinstallation or separate accounts

Route matching algorithm processing requests in under 2 seconds

Fare splitting engine with transparent per-rider cost breakdowns

Real-time ride tracking with ETA updates every 10 seconds

Features Delivered

Dual Driver/Passenger Mode

Smart Route Matching

Fare Splitting Engine

In-app Chat & Calling

Have a Similar Project?

We can discuss how our experience applies to your specific requirements.

Other Case Studies

EdTech

AnzanPro — Building a Multi-Platform Digital Abacus Learning System

Traditional abacus learning relies on physical soroban manipulation, making remote instruction difficult. The core engineering challenge was replicating realistic bead physics and multi-touch gesture recognition across Android, iOS, and Web from a single codebase — while maintaining sub-100ms touch response to feel tactile rather than laggy.

PYQBank — Adaptive Exam Preparation with Spaced Repetition

NEET and JEE aspirants need structured access to 22,000+ previous year questions organized by chapter, topic, and difficulty — with a study system that adapts to individual retention patterns. Existing platforms offered static question banks without intelligent review scheduling or exam pattern analysis.

VegPick — Three-App Grocery Delivery Ecosystem

Building a grocery delivery platform requires three distinct applications — customer storefront, vendor inventory management, and rider delivery tracking — that must operate in real-time coordination. The challenge was maintaining shared business logic across apps while handling concurrent order state transitions, live GPS tracking, and payment reconciliation.