Fleet Signal

A fleet operator runs on radio calls and guesswork: dispatchers phoning drivers for locations, fuel costs climbing with no explanation, maintenance happening after breakdowns instead of before them, and customers calling to ask where their delivery is.

An event platform for metal and rubber: GPS trackers stream positions through an ingestion service into a time-series store; geofences and rule engines turn raw pings into operational events — arrivals, departures, dwell, deviation, fuel anomalies. Dispatchers work a live console; customers get a tracking link; the fleet manager gets tomorrow's maintenance list.

• ▸MQTT/HTTP ingestion service, horizontally scaled, buffer-tolerant
• ▸Time-series store for positions & telemetry; PostgreSQL for entities
• ▸Rules engine for geofences, deviation and anomaly events
• ▸Alerting with escalation paths (console, SMS, email)
• ▸Public tracking links with scoped, expiring tokens

A dispatcher's console that answers three questions at a glance: where is everything, what is wrong right now, what needs deciding. Vehicle detail pages replay any day as a timeline — trips, stops, fuel, events.

Anomaly detection on fuel-versus-distance signatures; route deviation alerts tuned to reduce noise; maintenance hints from duty cycles and fault-code history — suggestions to a human planner, never silent actions.

Targets are what this design is dimensioned for — not results claimed from work we haven't done. We think you should hold us to exactly this standard of honesty.

• Design target — 30-second event latency from vehicle to console
• Design target — offline buffering with clean catch-up on reconnect
• Design target — alert precision tuned to under 5% false positives