The client
SensorMine is an Australian industrial intelligence company. Its platform connects assets, data and people in one AI-enabled system for industries such as mining, rail, water utilities, agriculture and transport, from condition monitoring across a mining fleet to real-time flood monitoring across river catchments and stormwater infrastructure.
The challenge
Many of the assets that matter most sit where there’s no power, no network and nobody on site. Crews drive for hours to check equipment that is usually fine, and occasionally has already failed. SensorMine needed hardware tough enough for those sites, and software that turns what it measures into a handful of decisions rather than a flood of threshold alarms.
What we built
- Nexus edge device. A rugged, industrial-grade device (IP67, −40 °C to +85 °C) that is commissioned in about five minutes. It connects analogue, digital, 4–20 mA, Modbus RS-485, RS-232 and 1-Wire sensors, runs on battery (up to 10 years), solar or mains power, and sends data over LoRaWAN, NB-IoT or satellite.
- Edge-AI camera systems. Three form factors on a shared NVIDIA Jetson platform, from a solar surveillance trailer that deploys in about 30 minutes to permanent fixed installations. On-device inference classifies safety events and anomalies in under two seconds, without streaming footage to the cloud.
- The SensorMine platform. A single device manager across more than 200 protocols that sits on top of existing systems, with dashboards, alerts and AI that learns each asset’s normal baseline and surfaces only what matters.
- AI agents over MCP. Every capability is published over the Model Context Protocol, so teams can query their operations from the AI tools they already use, while autopilot agents watch sites around the clock, investigate anomalies and draft work orders.
The outcome
SensorMine can put sensors and cameras on remote sites in minutes, keep them running without mains power or a network connection, and give operations teams a small number of meaningful alerts, all through one platform from the edge device to the dashboard.
