JaboToto: Building an E-Rickshaw Aggregator from Scratch

Kolkata had a clear gap in the mobility market: E-Rickshaws and bikes were everywhere, cheap, and practical for short urban trips — but they had no technological infrastructure to operate at scale. Existing ride-hailing platforms ignored this segment entirely.

Building JaboToto meant solving everything from scratch: real-time driver-passenger matching, geospatial search, a custom routing engine suited to Kolkata's streets, and a scalable backend that could grow as the driver and rider base expanded — all with no prior codebase to build on.

I designed and built the full platform as the founding CTO:

• Go Architecture: Built both a monolith and supporting microservices in Go, using WebSockets for real-time bidirectional communication between riders and drivers.
• Pub-Sub for Throughput: Implemented a Pub-Sub messaging layer (MQTT + event queues) to decouple ride requests and driver updates at scale, handling high message volumes without bottlenecks.
• Geospatial Stack: Used Redis for geospatial indexing and driver location caching, PostgreSQL as the primary data store, and Amazon S3 for static assets.
• Custom Routing Engine: Integrated OSM (OpenStreetMaps) ORS as the routing engine — giving accurate local routing without dependency on expensive third-party map APIs.
• In-house Search: Built a location-based search engine using Elasticsearch on top of internal location data, enabling fast driver discovery by area.

JaboToto launched and grew to meaningful scale before I handed operations to a trusted associate and moved on:

• 1,000+ Drivers Onboarded: The platform successfully onboarded and managed a large driver network across Kolkata.
• 5,000+ Customers: Reached a substantial rider base, validating the product-market fit for affordable urban mobility.
• Operational Handoff: After achieving growth, I transitioned operations smoothly to a trusted associate — a deliberate exit, not a failure.
• Technical Proof: Demonstrated that a lean Go-native architecture with open-source routing could power a full ride-hailing product at real-world scale.