Luma Energy
Student Information System Development for universities
800k meters streaming in near real time
Engineering
Connected products that keep working when the network does not.
Firmware, connectivity, streaming ingestion and analytics from one team. Our platforms run fleets from hundred-unit pilots to 800,000 smart meters.
Overview
The hard problems in IoT are at the edges: weak cellular coverage, hardware that must last ten years in the field, and fleets much bigger than the teams that look after them. We plan for this from the first diagram. That means offline-first firmware with store-and-forward buffering, ingestion that expects late and duplicate data, and a separate identity for every device.
We treat a device fleet like any other distributed system: it needs monitoring, automated deployment and failure budgets. That means over-the-air (OTA) updates with staged rollouts and automatic rollback, health data for every device, and a provisioning process that works at factory scale. A connected product that cannot be updated in the field is a recall waiting to happen.
The data side is where IoT projects succeed or fail. Telemetry is only worth the connectivity bill if it changes decisions, so we build the streaming pipelines, time-series storage and analytics together with the devices. The result is outage detection, anomaly alerts and forecasts that show up in an operator's daily work, not in a dashboard nobody opens.
Capabilities
Firmware, connectivity, ingestion and analytics, built together so the devices and the platform fit each other.
Firmware and edge software
Embedded development in C, Rust and Zephyr RTOS for limited hardware, with edge processing that filters and combines data before it ever reaches the network.
Resilient connectivity
MQTT, LoRaWAN and LTE-M, chosen to fit the power and coverage of your site, with store-and-forward buffering that treats dead zones as part of the design.
Streaming ingestion at scale
Pipelines built on Kafka and Kinesis, with Flink stream processing, that handle billions of events a day. They are built for late, duplicate and out-of-order data from day one.
Fleet management and OTA
Over-the-air updates with staged rollouts, automatic rollback and health monitoring for each device. A bad release reaches a few devices, not the whole fleet.
Device identity and security
Hardware-backed identity for each device, mutual TLS, and automated certificate renewal, so one compromised device never gives access to the rest of the fleet.
Time-series analytics
TimescaleDB and purpose-built stores with continuous aggregates, feeding anomaly detection and forecasts that reach operators inside the tools they already use. Grafana dashboards are there where a dashboard is the right tool.
Process
No black boxes. Each phase has explicit outputs, and you see progress weekly, in staging, not slide decks.
- 01
Hardware and protocol audit
We list the devices, radios and protocols you have or plan to buy, and we test the vendors' claims about coverage, power use and message costs.
Outputs
- Hardware audit
- Protocol assessment
- Message cost model
- 02
Connectivity architecture
We choose the protocols, design the message schemas, and agree the offline-behavior contract: what the device does when the network drops, and how the platform catches up when it returns.
Outputs
- Message schemas
- Offline-behavior contract
- Architecture doc
- 03
Pilot fleet
A production-grade pilot of 50 to 500 devices in real field conditions, not a lab demo. We add instrumentation to find the failures that only real locations can cause.
Outputs
- Pilot fleet report
- Field failure log
- Telemetry dashboards
- 04
Scale-out and hardening
We load-test ingestion beyond your fleet size, practice OTA updates including the rollback, and review security across the device, transport and platform layers.
Outputs
- Load-test results
- OTA rollback rehearsal
- Security review
- 05
Fleet operations handover
Runbooks, dashboards and training so your operations team can run the fleet day to day. Our engineers stay on call behind them for as long as you want the backup.
Outputs
- Runbooks
- Ops dashboards
- Training sessions
Outcomes
The results this service is accountable for, the reasons clients keep the engagement running.
No data loss offline
Devices store data through network outages and sync cleanly afterward, so gaps in coverage never mean lost data.
Proven at scale
We test your ingestion pipeline beyond your fleet size before launch, so you do not find its limits afterward.
Fixes without truck rolls
Over-the-air updates with staged rollout and rollback let you fix problems in the field in hours.
Contained risk
If one device is compromised, the damage stops at that device.
Data that drives action
Telemetry feeds daily operations such as alerts, work queues and forecasts.
A fleet you can run
Your own operations team gets runbooks and training as part of the project.
Stack
Chosen for your constraints, not our comfort, this is the toolkit we know deeply for this kind of work.
- Languages
- C
- Rust
- Embedded & IoT
- Zephyr RTOS
- MQTT
- LoRaWAN
- AWS IoT Core
- Data & streaming
- Apache Kafka
- Amazon Kinesis
- Apache Flink
- TimescaleDB
- Cloud & infrastructure
- Kubernetes
- Delivery & observability
- Grafana
FAQ
How we work with existing hardware, handle poor connectivity and price IoT projects.