Services

Our clients rarely start from a blank slate. They have grown system landscapes, processes with many stakeholders and deadlines that cannot be moved. We build software that fits into this landscape and then run it ourselves. You build it, you run it: that keeps quality high.

Custom software for complex processes

Ordering platforms, business applications, self-service portals: we develop applications that sales, back office and partners work with every day. Typically there are many product variants, approvals, documents and interfaces. Instead of customising standard software until nobody dares to install updates any more, we build exactly what your process needs and keep it up to date.

• Java 21/25
• Spring Boot 4
• Vaadin 25
• React & TypeScript
• PostgreSQL
• MySQL

UI/UX design

Good business applications do not need a training manual. Our UIs are so simple that training becomes largely unnecessary: new colleagues just get started instead of fighting their way through screens with 40 tabs and cryptic transaction codes. Together with specialised design partners we design user interfaces that are tailored to your users and their workflows and explain themselves. Figma designs and clickable prototypes make every feature tangible before it is commissioned. Our developers then implement the designs directly, without friction between agency and implementation.

For each client we maintain a dedicated component library: built in advance, documented in Storybook, visually tested and reused for every new use case. New features look consistent from day one and go live faster, because nobody reinvents the wheel.

• UX concepts
• UI design
• Figma
• Clickable prototypes
• Component library
• Design systems
• Tailwind CSS
• Storybook

Workflow automation and integration

An order runs through five systems and three departments and needs a signature at the end? We model such processes in BPMN, automate what can be automated and make the rest transparent. We connect to whatever is already in place: SOAP services and enterprise service buses as well as modern REST APIs.

No order gets lost along the way: every job either completes without errors, goes into retry on temporary failures or finally surfaces as an incident, with all the data needed to resolve it. There are no silent error rates quietly eroding service quality.

• BPMN with CIB seven / Camunda 7
• Retry & incident handling
• API-first with OpenAPI
• DocuSign
• Microsoft 365 (Graph)
• Keycloak / OAuth2
• OpenSearch

Product modelling and rule engines

Ordering a pizza is easy, configuring a car less so. A business connection with fibre, backup line and service level is in another league altogether: bandwidth, router, contract term, availability at the address, building costs and discounts all depend on each other, and the rules change with every sales release.

There is nothing off the shelf for products like these. Commercial CPQ systems and catalogue standards such as TM Forum TMF620 describe products and prices, but not the logic behind them. We have designed and built a product catalogue with a rule engine for a major telecommunications provider, so we know what matters:

• Declarative rules for dependencies, exclusions and calculations, evaluated live during configuration
• Context instead of static price lists: address, roll-out status, building costs, carrier quotes and customer account feed into price and approval
• Price corridors with term-dependent floors; going below them automatically triggers an approval
• Technical order data is captured and validated during the sale, so nothing has to be clarified later in fulfilment
• Rules tested like software: automated regression tests per product, so a change does not silently break other products
• Versioning, so open quotes, orders and contracts keep the product version they were created with

The hardest part is rarely the code, it is understanding the product: before a product can be represented digitally, it has to be understood. We translate sales releases, price sheets and the knowledge of product managers into testable rules and keep them up to date in live operation.

If your products are similarly complex, we build you a solution that reflects exactly your rules. We bring the know-how of where such projects typically fail.

• Rule engines
• Declarative rules
• Regression tests per product
• BPMN approvals
• TM Forum TMF620

Cloud and infrastructure as code

We build environments that can be reproduced at any time: entirely as infrastructure as code, separated by environment and account, with encryption, web application firewall and backups from the start. We move grown, hand-clicked setups into code step by step, without interrupting operations.

It does not have to be AWS: if your data has to stay in your own data centre, we build and run Kubernetes clusters and virtualised environments with Proxmox on standard hardware. Our own infrastructure in Hamburg runs on Proxmox as well.

• AWS
• ECS / Fargate
• Lambda
• Kubernetes
• Proxmox
• CDK for Terraform (TypeScript)
• Terraform / OpenTofu
• Ansible

Operations, monitoring and DevOps

Our developers prefer to run their software themselves. That includes automated pipelines, monitoring that reports problems before your customers do, and an on-call team that knows the system from the inside. We keep dependencies up to date continuously instead of setting up a big migration project every few years.

• OpenTelemetry
• Grafana, Prometheus, Loki, Tempo
• GitLab CI/CD
• Renovate
• Go for server tooling

Test automation

We test daily and continuously: unit and integration tests on every commit, plus real end-to-end tests through the frontend, either directly in the CI pipeline or separately against the target system. Visual regression tests catch unintended changes in content and design, and whatever cannot sensibly be automated is checked manually before a release goes live.

• Playwright
• Selenium
• Chromatic (visual regression)
• Storybook
• Testcontainers

Consulting and project rescue

Our speciality: firefighting. Project status amber or red, deadline at risk, the legacy system “can’t do that”? We quickly get the full picture, name the actual causes and get our hands dirty ourselves. From architecture reviews and performance analyses to the step-by-step replacement of legacy systems.

We go as deep as necessary: from the network stack and packet dumps in Wireshark to heap dumps, thread dumps and a debugger attached to the running system. In the end, the root cause is found in the code or in a packet dump, not in a presentation.

How we work

• Curious by principle: We love new subject areas. Some of us are engineers, and the building continues after hours: solar installations, smart homes with KNX and Home Assistant, firmware for irrigation systems. We bring this curiosity to every project, and the thoroughness with which we evaluate new things as well.
• Compact, experienced, direct: 25 people in Düsseldorf and at our subsidiary in Prague. You talk directly to the developers, without a cascade of project managers.
• Pragmatic, not hype-driven: We evaluate new technologies thoroughly in-house. Whatever makes it into your project has proven itself.
• AI-assisted, with judgement: We use AI coding agents in our daily development, with clear project rules, code reviews and automated tests. It makes us faster, not sloppier.
• Truly iterative: We work with Scrum or Kanban and usually deliver a release with finished, working features every 14 days. Time to market stays short and you see early on what you are getting.

Sounds like your problem? Get in touch.