Industrial Systems Architect · OT/IT Integration · Complex System Diagnosis
When systems stop talking, I find out why.
The problem usually sits between two systems. That's where I work.
Stubborn technical problems rarely live inside one system. They sit between two: the PLC and the network, the database and the integration, the software and the way people actually use it. That's why nobody finds them.
I'm Pieter Leek, an independent systems architect. I step into stalled integrations, unclear OT/IT environments and projects drifting toward go-live. I map what is actually happening, replace assumptions with evidence, and fix or build what's needed. So management can decide and teams can move again.
Built for problems without a manual.
Where I Help
Diagnosis, integration and recovery for environments where systems, people and decisions no longer line up.
Complex System Diagnosis
Something keeps failing and the fix never holds. Or everyone has a theory and three vendors point at each other. I read the whole chain: field devices, PLCs, network, middleware, data, software and the people using it. The cause usually sits across two of those layers. I trace it, explain it in plain language, and make sure it doesn't come back.
Typical: intermittent data loss between the floor and SCADA or MES · a network that works until it doesn't · alarms nobody trusts anymore.
OT/IT Integration, Legacy & Middleware
Machines, PLCs, databases and cloud platforms that don't share a language. I design and build the layer in between: protocol translation, edge data services, API wrappers and message pipelines. Documented, so your own team can own it afterwards.
Typical: getting data out of a 15-year-old production line · connecting OT data to ERP or cloud without exposing the plant network.
Sensor-to-Cloud Architecture
From field sensors and edge hardware to storage, dashboards and alerts. I design power, connectivity, data model and security as one system. Built to survive the field, not just the demo.
Technical Recovery & Go-Live Support
Go-live is close and the project is drifting, or the environment has become too unclear to steer. I come in fast, separate symptoms from causes and define what must be true before go-live. Then I help the team get there.
Architecture Between Boardroom and Floor
Management needs a direction it can trust; engineers need a decision. I translate the technical reality into clear options, each with its cost and risk. The decision gets made on how the environment actually works.
How an Engagement Works
Short, intense and with a clear end. You always know what you get next.
- Send the context. A few lines on what's failing and what has been tried. I reply personally.
- Intake call. 30 minutes, free. We decide whether I'm the right person and what the first step is.
- Diagnosis. Typically two weeks. I map the system, find the real constraint and deliver a written findings report with options.
- Recovery or build. Typically three to six months. I fix or build what's needed, working alongside your team.
- Clean handover. Documentation, knowledge transfer, and then I step out. I build solutions; I don't stay for maintenance.
Remote-first by design. On-site when the hardware, the integration or the situation requires it.
Selected Work
A selection of systems I designed and delivered for clients. Names withheld where agreements require it.
De-icing Control System
Designed and built a control system for de-icing, from sensors and actuators through control logic to the operator interface.
Remote Monitoring for Neonatal Incubators
Designed the chain from incubator to remote monitoring, where a missed reading is not an option: data capture, transport and alerting.
Cashless NFC Payment for a Festival
Built the full system for a festival of about 600 visitors: NFC wristbands, point-of-sale, and a backend in Node.js/TypeScript and Java Spring Boot.
International E-commerce and Supply-Chain Platform
Built the platform and integrations for a company running sales, ERP and manufacturing in China: ERP, factory workflow, a credit-insurance API (Atradius) and a FileMaker backend. Implemented on-site in Chicago.
Current Work & Prototypes
What I'm building and running now. It keeps my hands on real hardware and real users.
Amerta One: Solar-Powered Salinity Monitoring for Rice Fields
A compact field module that measures salinity stress risk in rice fields. Farmers see a simple LED status; the data goes to Tirta Cloud for research, calibration and future advice. It combines ESP32/LoRa hardware, solar and LiFePO4 power design, and cloud data, built for a muddy field in Central Java rather than a demo table.
Status: prototypes in development, field testing planned in Java. → Visit amerta.one
Multi-User Platform with AI Feedback
Designed and run a platform for 150+ concurrent users: 30 isolated Docker containers behind a Caddy reverse proxy, with a custom Python backend that evaluates user work with AI.
Compliance Feedback Engine
Checks unstructured user input against a formal regulatory framework and returns structured feedback within seconds.
MQTT Relay Bridge
Turns a dashboard command into a physical action: MQTT messages switch hardware relays on analog equipment. A small build, but the same pattern as connecting legacy equipment to a modern control layer.
Writing
How I think about industrial systems, written down.
Designing Defensible Industrial Systems
A target architecture for OT/IT environments that can withstand attack and still be operated. It works through segmentation, OT firewalling, certificate-based device identity and secure telemetry, aligned with IEC 62443, NIS2 and ISA-95. Built on industrial-grade Siemens hardware, with a companion video series on Zero Trust in OT/IT, in English and Dutch.
→ Download the paper (PDF, v1.1)
Notes
Shorter pieces on system boundaries, integration and complexity. → Read the notes
About
I work where technology, people and decisions meet, especially when the situation no longer fits a manual.
I've been programming since 1982 and was on the Dutch internet in 1991. I trained in electronics, built software and infrastructure for international clients, and led a software engineering team of 25. Along the way I chaired two boards, through a €450k harbour renewal and a merger of three organisations.
I started out in community development. That's where I learned that a system only works when the people around it can work with it. It's why I move as easily between the shop floor and the boardroom as between software stacks.
That breadth is deliberate. Most stalled projects need someone who can follow a problem across hardware, network, software and organisation without handing it off. I use the part of the toolkit the situation needs.
The name Qman is a nod to Q from James Bond: the quartermaster who stays out of frontline politics, works from his own lab, and builds exactly the hardware and software needed to survive the mission.
Toolkit
Something stalled, failing or unclear?
Send me a few lines of context: what's going wrong, what has been tried, and when it needs to work. I reply personally.
From sensor to cloud, from shop floor to boardroom.