My own projects, work from my apprenticeship and studies, and the companies I've worked for. Each entry says what it is, what I did and what I used.
A Flutter monorepo as the starting point for my own apps
Each new app copies the template once and then goes its own way. It includes my own MVC state management with a lint plugin that enforces its rules, localization and a base setup for macOS, Windows and Linux. It also has Dart packages of my own whose code is generated with build_runner. barrierewacht is being built on it.
A tool web agencies use to check their clients' websites for accessibility problems
An agency adds the websites it looks after. barrierewacht checks them automatically against WCAG 2.2 AA, ranks the problems by importance and checks again on a schedule. It's an automated audit tool, not a certified BITV test and not legal advice.
Workflows and working principles for Claude Code, kept in markdown files under version control
groundwork keeps what you learn about working with AI coding agents in markdown files: workflows, working principles and slash commands. That puts it under version control instead of in chat history. A small Python script assembles the files into the directory Claude Code reads at the start of every session. It also reports notes that have drifted, such as a broken link. I published groundwork as open source. Your own details, such as the vendors you use, go in a private layer on top.
Minecraft mods in Java, with public source code
Work on a remote maintenance platform and three Unity apps
At oculavis I continued my apprenticeship and joined the computer vision and augmented reality team. As an apprentice I worked on the main product, a platform for remote maintenance and remote support. Service teams and specialists use it to work on faults together. I also completed two white-label Unity apps for EU-funded research projects. Then there was another Unity app, built from a template, that shows machine data over MQTT; it was the topic of my seminar paper.
WPF user interfaces and internal research projects with OpenCV, in C#
As an apprentice I worked on the main application, on its WPF user interfaces in C#. In internal research projects with OpenCV, I detected a Siemens star in the camera image and computed the focus quality from it. I also wrote my own version of the focus peaking found in Sony cameras, which highlights sharp edges in the image. I helped build and install camera systems and ran tests on a near-production system in the test lab.
A Flutter app that links photos to the digital twin of a forest
For a required project at university, our group built a Flutter app. It was connected to the system of RWTH Aachen University's Institut für Mensch-Maschine-Interaktion (MMI), its institute for human-machine interaction. The institute was researching digital twins of forests, in the context of the research project “Kompetenzzentrum Wald und Holz 4.0”. You upload a photo, and our backend analyzes it and queries the digital twin server. The app then shows information on individual trees, staff or machines. We chose Flutter ourselves; my part was mainly the frontend. We presented the app at the university's software showcase.
Bringing machine data into a Unity app over MQTT
My seminar paper (title in German: „Einbindung einer MQTT Schnittstelle in Unity-Apps zur erweiterten Darstellung von Maschineninformationen“) shows how a Unity app receives and displays machine information over MQTT. The C# code excerpts are public: a client interface, a client built on MQTTnet and a connection manager. The topic came from my work at oculavis, from the Unity app for machine data.
At school I taught myself Java and built a timetable app for my own phone, which I used at school. Rebuilding it in Flutter was my reason to try Flutter. I never got around to it.