We build software and automation for businesses that need something specific done: a system that fits how the business actually works, not a template stretched to cover it. We start by understanding how things work today, agree on what the finished system needs to do, and only then start building.
The work splits into two capacity lanes (software and web on one side, industrial and AI automation on the other), and most projects sit mostly in one or draw a little from both.
Software and web development
Java. We build backend systems and APIs on Java, Spring Boot, and Hibernate. This is where the data model has to be right before anything else: account systems, internal tools, and services that other parts of a business rely on and that have to keep working quietly in the background. It is also usually the layer other systems get built on top of, so we treat getting it right the first time as the priority, not something to patch later.
Fullstack web. For projects that need a full path from interface to server, we work with Angular or React on the frontend, Node.js and TypeScript underneath, and PostgreSQL or MySQL to hold the data. We package and ship through Docker, host on AWS, and run the build pipeline through Jenkins, so a change goes from a commit to a live deployment through the same repeatable process every time, not a manual one-off. This is the lane most client web systems, booking tools, and internal dashboards fall into.
Industrial and automation systems
Siemens automation. We are certified in Siemens STEP7 and ladder-logic programming, with hands-on experience in Loxone smart-building systems. This is a smaller, specialist part of what we do, useful for businesses that need control-system or building-automation work done alongside their software, instead of bringing in a separate vendor for each. We are upfront about the scope of this: it covers certified fundamentals and hands-on building-automation work, not decades of heavy-industry plant integration.
AI automation. On the AI side, we work with Python and PyTorch, and with YOLO and DeepSORT for real-time video analysis and object tracking. In practice this means systems that watch a video feed continuously and count, detect, or flag something as it happens: not a one-off report, but automation that keeps running on its own. It is a good fit for businesses that currently rely on someone watching a screen or checking manually, and want that turned into a system instead.
Our own products
We also build and run our own products, and we treat them the same way we treat client work: scope it, build it, keep it running. They are not a separate business; they are proof of how we work, made visible.
Filigrelle is a laser-cut personalized gifts and home decor brand, sold through our own e-commerce shop. We built the storefront and the order and fulfillment systems behind it, and we still run it day to day: the same fullstack and automation work described above, applied to our own product instead of a client’s.
PraxAI is our practical AI training, audit, and automation consulting practice, for businesses that want to use AI in their day-to-day work rather than just talk about it. It draws on the same AI-automation capability (model work, video and data analysis) but packaged as guidance and hands-on implementation for another business, instead of a product we run ourselves.
See both, and the kind of client work we can describe without naming names, in more detail on our Portfolio page.
How a project starts
A project usually starts with a short conversation about what you need and why the current way of doing it does not work well enough. From there we write down the scope before any work begins, so both sides agree on what “done” looks like, including what is out of scope, so there are no surprises later. Then we build, with plain-language updates as we go rather than silence until a deadline. After launch, we stay reachable: the person who built the system is still the person you can reach if something needs to change. And if a project does not fit our current capacity or the areas above, we say so early, rather than take it on and figure it out as we go.
Have a project in mind?