How We Design an IoT Monitoring System

Our process for turning a monitoring requirement into a working IoT system, from sensor selection to a cloud dashboard.

Start with the constraint, not the sensor

Every IoT monitoring project we take on starts with one question: what has to be true in the field for this device to work? For T-Remo, our temperature monitoring device for vaccine and medicine transport, that constraint was operating reliably at negative temperatures inside a moving vehicle with intermittent connectivity. That single constraint shaped the sensor choice, the enclosure, the power budget, and how often the device could safely transmit.

We treat the environment as the spec. A warehouse monitoring project and a cold-chain transport project can use similar sensors but need completely different power, connectivity, and enclosure decisions.

Choosing the sensor and the edge device

We build on ESP32 and Arduino, Raspberry Pi, STM32, LoRa, and NRF32, choosing the platform based on processing needs, power budget, and how much logic needs to run on the device itself versus in the cloud.

For a device like Tracker, which is battery-operated and needs to run unattended for extended periods, the priority is power efficiency and onboard aggregation — the device should only transmit meaningful, aggregated data rather than a constant raw stream.

Getting data off the device reliably

Connectivity choice depends on where the device physically lives. A fixed installation with existing infrastructure can use Wi-Fi. A device that moves, like a cold-chain shipment, needs something that works without relying on a facility's network — which is where cellular (SMS/data) or LoRa come in, depending on range and power constraints.

What happens after the data arrives

Once data reaches the cloud, it needs to become something a person can act on: a live dashboard, a threshold-based alert, or a report. Our IoT products push data to cloud dashboards and can trigger remote alerts through SMS, email, or app notifications when a reading crosses a defined limit.

Testing in the real environment

A monitoring system that works on a desk and fails in a delivery van hasn't been tested properly. We validate devices under the actual conditions they'll run in — temperature extremes, motion, and connectivity dropouts — before they go into production use.

Related products