Marek Rjelka

OT/IT Integration Sprint · 6 weeks · fixed fee

Your machines in a dashboard your ops team trusts. In six weeks. Fixed fee.

OPC UA, MQTT, Modbus and proprietary PLCs into a time-series store, Grafana boards and alerts. Runs on your hardware, in your Docker. No SaaS lock-in. No twelve-month systems-integrator project.

Book a 30-min discovery call → See the live demo ↗

The problem

The offer

A fixed-scope, fixed-fee six-week engagement that turns your existing machine signals into a working data layer: gateway, time-series database, Grafana dashboards, and alerting. You keep the source code. It runs on your hardware, in your Docker.

Six weeks, in detail

  1. Week 1

    Audit, tag list, KPI workshop

    • On-site protocol audit: what each machine actually exposes
    • KPI workshop: which numbers will move the business
  2. Weeks 2–4

    Gateway, ingestion, dashboards

    • OPC UA / MQTT / Modbus gateway deployed in your Docker
    • Database ingestion with retention policies
    • Grafana boards built against the agreed KPIs
  3. Weeks 5–6

    Alerts, runbook, handover

    • Alerting rules wired to your existing channel (email, Slack, Teams)
    • Operations runbook and on-call response guide
    • Hands-on training for ops and IT, no black boxes
Book a 30-min discovery call →

Proof

The exact stack used in this sprint runs as a public demo:

factory.rjelka.de ↗

OPC UA MQTT InfluxDB Grafana Docker

Past industrial clients include:

Is this for you?

A fit if

  • 5–50 machines, single plant or pilot line
  • You have an internal IT or ops team that can host Docker
  • Your machines speak OPC UA, MQTT, Modbus, or have an exposed API
  • You want to own the stack, not lease it

Not a fit if

  • You've already standardised on Ignition, MindSphere or AWS IoT: call them, not me
  • You need real-time control loops under 10 ms. This is observability, not control.
  • You have zero IT capacity to host the deliverable post-handover
  • You expect AI to magically fix your data quality. Garbage in still equals garbage out.

Frequently asked

Why fixed fee and not time and materials?

Because you should know what you're getting. The scope is one plant or one line, your specified protocols, and the deliverables on this page. If something falls outside that scope we discuss it before adding it, no surprise invoices.

What if our protocols are exotic or proprietary?

Most 'exotic' protocols are TCP/XML, Modbus or a vendor REST API in disguise. I have integrated MPERIA, Matthews, custom XML-over-TCP printers and PLC/scanner combinations at SpaceX and Volkswagen. The week-1 audit confirms feasibility before we commit.

Who owns the code afterwards?

You do. Source code, configurations and dashboards are delivered to your repository. No license, no escrow, no vendor lock-in.

Can it run on-premises and air-gapped?

Yes. The default deployment is on-prem in Docker. Air-gapped installations add a few days for offline image distribution but are well-trodden ground.

What about long-term support?

After handover I offer optional retainer support for changes, additional lines or new dashboards. Most customers don't need it after month two, that's the point.

Why isn't this just MindSphere or Ignition?

If you've already invested in MindSphere, Ignition or AWS IoT, keep going. This sprint exists for plants that don't want to lock their data into a vendor pricing model, or have tried one and got stuck.

What does week 0 look like, before we start?

30-minute discovery call → fit assessment → fixed-fee proposal with scope and protocols. Typically 1–2 weeks from first call to signed scope. The on-site audit (week 1) starts after that.

What's not included?

PLC programming, MES integration, ERP integration, custom hardware, control logic, and anything that requires a functional-safety review. Those are separate engagements.

Talk in 30 minutes

Book a 30-min discovery call →

Free, no slides, no sales deck. We discuss your plant, protocols and goals, you leave with a clear sense of fit.