IIoT platform engineering - telemetry, device fleets and the systems built on them, at production scale

Where we fit

Sensors and controllers are the visible part of an IIoT system. What decides whether it delivers value is the software behind them: how device data is ingested, stored, made queryable, secured, and turned into something an operator or a customer uses. That layer has to keep working for the lifetime of the hardware, which is far longer than a typical software project.

What we have actually built

Connected HVAC, Ariston Group (Wolf) - 100,000 devices. We modernised a legacy SaaS IoT platform that could only retain device data for 24 hours. We added an MQTT layer bridging the existing TLS server to a long-term data platform, restructured core components to run multi-instance behind a load balancer, and moved customer tools, mobile apps and back-office systems onto the new platform. In a second engagement we handled project management and integration for a new MQTT backend on bare-metal Kubernetes using EMQX, with GraphQL for server-to-server traffic and custom microservices deployed in the customer's cloud. Around 40 GB of traffic per hour against 80 TB of stored data.

Connected fleet, BMW - 10,000+ vehicles, 20,000 active connections. A vehicle communication platform for car-sharing: status and location through back-office tools, and the same data surfaced to customers in the mobile app. Built on a scalable HiveMQ MQTT broker supporting both synchronous and asynchronous exchange, with purpose-built microservices for message handling and orchestration, delivered across seven coordinated teams.

Medical equipment, SciCan - 10,000 devices. A monolithic application rebuilt as a high-availability architecture on AWS IoT Core, with real-time monitoring and control, after an explicit cost-versus-volume analysis of the managed-service option.

Manufacturing park connectivity - 30+ machines. Machines with no network presence brought online with an on-premise MQTT server, working alongside the customer's production development team on module selection and installation, with Ignition and custom services for analytics and process automation.

Also: Siemens cloud and controls end-to-end integration (proof of concept, objectives met); MQTT-connected laboratory machines and remote machine control for TH-OWL; real-time status, location and event monitoring for vertical transport equipment.

Technical ground we cover

• Brokers and streaming: HiveMQ, EMQX, Kafka, ActiveMQ, MQTT generally, RTP

• Deployment: public cloud, bare-metal Kubernetes, hybrid and fully on-premise

• Device and protocol layer: CANbus, RS232, I2C, I2S, Bluetooth, USB, custom drivers, embedded software

• Architecture: event-based, queue-based and stateless REST; asynchronous backlog processing; architectures scaled to millions of requests per second

• Integration: GraphQL and REST APIs into MES, ERP and reporting systems; low-code layers where a visualisation or management surface is needed quickly

How this differs from a typical development contract

We build systems we expect to operate. Everything is instrumented for monitoring from the start, deployable through automated pipelines, backed up and restorable, with identity handled centrally. A shortcut taken during the build becomes our incident at 3am.

We do not design or manufacture hardware, and we do not compete on hourly rate.

What are we looking for?

Commercial agreement with technical assistance, subcontracting, and R&D or innovation project partnership. We are looking for device manufacturers, industrial automation and energy companies, and research organisations that have the hardware and the domain expertise but need a long-term software and platform partner.