HomeServicesTechnical leadership and the platform underneath
Technical leadership and the platform underneath
Many mid-sized companies cannot get their machine data off the shop floor, and once they do, the environment that would make the data worth something is missing. We build that environment and take on technical leadership where nobody in house can carry it.
Getting data off the shop floor
Machines speak OPC UA, but that alone puts nothing where it is needed. Whether the route runs through an MQTT gateway or straight onto an event bus depends on the plant and on who is going to operate it afterwards.
Turning readings into information
192 degrees Celsius is not a statement. Only with a defined range, a warning threshold, a link to the batch and the right aggregation does it become something anyone can act on. Those rules are written with your people from production, not at a desk.
Connecting old and new
Almost every plant has a system that has to stay. We connect it rather than work around it, and settle beforehand which side holds the truth.
The platform it all runs on
Containers, registry, sign-in, distribution, monitoring. The work nobody sees, and without which nothing moves in year two.
Reviewing instead of building
Sometimes you do not need us as the builder but as a second opinion. We read architecture proposals and vendor quotes and tell you what is missing from them.
Two ways to bring us in
Both roles draw on the same knowledge but cost very different amounts.
As a second opinion
You have a quote, an architecture proposal or a running project and want to know whether it holds. We read along, ask the questions nobody asks internally, and write down what we found. It is the smallest way in and often the most useful.
As technical leadership for a period
Where nobody in house can own the architecture decisions, we take the role for an agreed period. We do not decide past you; we prepare decisions so that your management can take them.
What we built and ran there
Since 2017 we have provided part of the technical leadership at a manufacturing-software vendor. This is the stack we are responsible for.
Runtime
Tools
Kubernetes on AWS EKS, Helm, Consul for service registration
Running a cluster is not setting one up; it is version upgrades in production, networking trouble on individual nodes, and the question of where persistent data lives when a node disappears.
Distribution and registry
Tools
Artifactory as the registry, migration onto S3, JFrog Xray for scanning
Anyone shipping containers needs one place where it is traceable what an image contains. Migrating a grown registry without downtime is work in its own right.
Security and evidence
Tools
SBOM management, CVE handling in the pipeline, hardened base images with a read-only root filesystem
Manufacturing software ends up in environments where somebody has to prove what was shipped. A scanner alone does not do that; the findings have to be worked off.
Sign-in and permissions
Tools
Keycloak, OAuth and OpenID Connect, connection to existing directories
In a factory, permissions hang off plant, line and shift. The permission model is rarely a technical problem; the question is who owns it on the business side.
Data flow
Tools
Kafka as the event bus, connection of existing systems, modelling along ISA-95
For a reading from the plant to arrive in a control room, it has to be established on the way which equipment and which order it belongs to.
Large vendors bring a fixed team for this and invoice accordingly. We come with as many people as the task needs, sometimes with one. What we bring is the mistakes from many years of running such a platform, and an early word on which parts you do not need at all.
First the conversation, then the analysis, then the quote.
Tell us what you run and what gets in the way. If it fits, we look at your system more closely. Whether that analysis runs as a separate service or feeds into a quote, we tell you before we start. After that you get an assessment: effort, timeframe, price range and the risks we see.
Discuss your project