Connected operations
Engineering a multi-tenant fleet intelligence platform
A cloud-native platform that turns live vehicle telemetry into secure, role-aware operational workflows for fleet operators and service providers.
The context
The operating problem we designed around
The product had to ingest high-volume GPS data from varied devices while supporting responsive dashboards, APIs, background processing, and real-time updates. Operational users needed clarity during active fleet events, not another disconnected reporting tool.
The architecture also needed to accommodate organisations, sub-accounts, configurable modules, permissions, billing, and future product expansion without coupling every capability to the telemetry pipeline.
Fleet data is only valuable when teams can trust it, understand it, and act on it. The platform needed to support live tracking, historical analysis, alerts, maintenance, reporting, and white-label operations while keeping each tenant securely isolated.
Delivery approach
From ambiguity to an operable system
Model the operating hierarchy
Defined accounts, organisations, users, roles, fleets, vehicles, devices, and entitlements before building feature workflows.
Separate ingestion from experience
Isolated protocol parsing, telemetry persistence, event processing, APIs, and user-facing applications behind clear boundaries.
Deliver in operational slices
Built complete paths—device event to operator action—then expanded into history, alerts, reports, and maintenance.
Engineer for operability
Added health checks, structured logs, retention policies, queues, and deployment controls as platform capabilities.
The solution
Capabilities designed as one operational system
Live operations workspace
Real-time maps, vehicle state, geofences, alerts, assignments, and role-specific fleet visibility.
History and intelligence
Trips, routes, stops, playback, reports, exports, and analytics backed by purpose-designed telemetry storage.
Fleet administration
Vehicles, drivers, devices, documents, maintenance, fuel, expenses, users, and organisational controls.
Platform extensibility
REST APIs, webhooks, module entitlements, white-label foundations, and integration-ready domain services.
System architecture
Clear boundaries between experience, domain, and platform layers
LAYER 01
Device edge
TCP and UDP protocol services validate, normalise, batch, and preserve sampled raw packets for support.
LAYER 02
Operational data
PostgreSQL, PostGIS, and time-series storage support transactional, geospatial, and telemetry workloads.
LAYER 03
Application services
Go APIs, workers, queues, and real-time services isolate synchronous requests from event-driven processing.
LAYER 04
Experience layer
Responsive Next.js interfaces expose only authorised data and adapt workflows to enabled modules.
Advanced capabilities
Intelligence introduced with evidence, controls, and operability
Behaviour and risk signals
Derive harsh events, route deviations, excessive idling, and configurable operating exceptions from telemetry.
Predictive maintenance foundations
Combine utilisation, diagnostic signals, service schedules, and repair history for earlier maintenance decisions.
Operational assistants
Ground future AI assistance in authorised fleet data, documented workflows, and traceable source records.
Potential outcomes
What the engineering approach enabled
- One operational model across live tracking and fleet administration
- A scalable foundation for white-label and multi-tenant delivery
- Clear separation between device traffic and customer-facing workloads
- Modular growth across alerts, reporting, maintenance, and billing
Engineering lessons
What carries into the next platform
- 01Tenant isolation must shape the data model and authorisation path from the beginning.
- 02Telemetry retention and aggregation policies are product decisions, not only database decisions.
- 03Operator trust improves when live state can always be traced back to durable historical events.
Services applied
Technology landscape