IoT & Data Pipelines

The path from device to decision — protocols, edge software, ingestion and time-series data.

We build the path from device to decision: protocol handling at the device, edge software, ingestion, time-series storage and the interfaces on top. In production, systems we have built handle 340 million data points per day, with fleets connected over Modbus, EEBUS, MQTT and OCPP.

We take the data path as an area of responsibility — architecture, implementation, production behaviour and the monitoring around it — and work inside your team and your repository. Engagements run for months rather than sprints, and what we build stays maintainable by the people who stay.

Engagement model. Longer engagements, typically 3–12 months, part- or full-time. Remote or on site in the Munich area. German and English. We work inside your team and your repository, with responsibility for an area end to end.

Our Focus Areas in IoT & Data Pipelines

  • Device and plant connectivity
    Meters, sensors, inverters, machines, energy assets and charging infrastructure over Modbus, EEBUS, MQTT and OCPP. Where a vendor system keeps control of the installation, we integrate alongside it — in one project our edge device ran in slave mode under Loxone.
  • Edge software
    Edge services on Linux hardware, including Raspberry Pi footprints: local buffering, store-and-forward across connection loss, and the control logic that has to run without a round trip to the cloud.
  • Fleet operation
    Secure remote access over Headscale and a tailnet, OTA updates, staged rollouts and fleet management, so releases and diagnostics do not require a site visit.
  • Ingestion
    Streaming and batch pipelines with Kafka and Apache Beam, batching for efficient ingest at volume, and long-running device workflows orchestrated with Temporal. Late, duplicate and malformed messages are handled explicitly, as is backfill.
  • Storage and interfaces
    InfluxDB and TimescaleDB for measurements, PostgreSQL for relational data, Redis for caching. Dashboards in Vue 3, and Flutter where the work happens in the field.
  • Control logic
    Peak shaving and load shifting against a site’s own load profile and available assets, and tariff-aware control — for example EV charging timed against dynamic tariffs and current load.

In production

  • B2B2C energy platform
    A white-label platform through which utilities and Stadtwerke give end customers access to their own consumption data. Device data collected over REST APIs and orchestrated with Temporal, time-series storage in InfluxDB, Apache Beam pipelines for batch ingest, backend in Go and Java/Spring Boot, frontend in Vue 3, mobile app in Flutter. 340 million data points per day; the platform grew from 700 to 20,000 users. Our role: solutions architect, from architecture through to production.
  • EMS with edge device for mid-sized industry
    Peak shaving and load shifting from the site’s own load profile. Edge services on a Linux-based Raspberry Pi, device data over Modbus and EEBUS, Loxone integration in slave mode, OTA updates and remote maintenance over Headscale and a tailnet, backend in Python/FastAPI, frontend in Vue 3.

Technologies: Modbus · EEBUS · MQTT · OCPP · Loxone · Linux / Raspberry Pi edge · Headscale / Tailscale · Kafka · Apache Beam · Temporal · InfluxDB · TimescaleDB · PostgreSQL · Redis · Go · Java/Spring Boot · Python/FastAPI · Vue 3 · Flutter · Docker · Kubernetes · AWS · GCP

Send us the outline of your system.

Devices, protocols, volume, and what you need at the other end. You get a direct assessment of scope and fit.