The Godot game engine is a large open-source project with over 3 million lines of code. I contributed more than 100 isolated changes, with a focus to systematic improvements for program correctness and efficiency. I was eventually invited to be a maintainer in the 'Core' team, which designs, maintains and reviews change requests for low-level code and the overall program architecture.

The goal of the project is to facilitate scientific computation, popularized in Python by NumPy and SciPy, in Godot games. NumDot uses xtensor under the hood, and is SIMD accelerated with xsimd.

To address the lack of accessible ECG work solutions, I decided to dedicate about 2 weeks of my dissertation time to create a free, open source program with which ECGs and annotations can be viewed. To achieve this, I needed to combine several python development techniques to create a boilerplate with which dash-based apps can be published as standalone applications. I shared those efforts with the community as an open source project.

I created new versions of my personal websites (tenbrink.me, ivorius.de) with Astro.

I create a series of videos about game development. The videos are made to be entertaining and approachable, and thus require technical, artistic, and performative skills. For this project, I have also released several open source tools, such as a swift-based presentation toolkit, a code stepper and a Godot template for custom gravity.

While working in cardiology as a data analyst, I noticed that FAIR (findable, accessible, interoperable and reusable) code is rare in this field. I hypothesized that this may be partially addressed with a simple scientific programming language that can be used in all common data science environments. I started designing and implementing this language as a side project during my dissertation. At around 10,000 LOC, it features a VM, a python compile target, type inference, and compile-time code execution, among others. To make it practically useful though, a lot more work is required.

I am currently employed at LMU Klinikum in Munich, working on my doctoral thesis 'Robust automatic determination of T-Wave boundaries in ECGs'. I work closely together with medical doctors to analyze Electrocardiographic data, and develop new approaches to aid with cardiac diagnosis and risk assessment. My supervisors are Axel Loewe and Konstantinos Rizas.

I co-authored this patent during my employment at Ablacon, Inc. (now Cortex, Inc.). To isolate atrial signals, templates are created from mean heartbeats, and subtracted at heartbeat locations. This eliminates most of the signal synchronized to the QRS. In most cases, the most prominent residual is the atrial signal. To achieve satisfactory results, substantial research efforts with many data science techniques were needed.

My master's thesis at Technical University of Munich (TUM) involved working with 64-electrode data from real life patients. To collect the data, we collaborated with the LMU Klinikum in Munich. I collaborated on the techniques with my supervisors Philip Häusser and Peter Ruppersberg. The biggest challenge was the large variety of patient data, many of which required special attention. During this time, I also created a free open source library supporting complex GUI in python notebooks.

I designed, built and programmed a fan that can display 'holographic' images by rapidly toggling LEDs on the spinning blades. Since I started the project, this technology has become a popular gadget for advertising. My implementation is the only one that is public, free, and can stream live image data from a server. This project involved designing a CAD model to be 3D printed, custom PCBs, sourcing parts, wiring and soldering, and the software for the client and the server of the device. I also stripped down the LED driver part of the project to share it with the public.

I worked at Ablacon, Inc. (now Cortex, Inc.), first as a student, then full time. During this time, I contributed to their cardiographic ablation software as a software developer. I also set up a cloud computation pipeline in aws with ray.

My bachelor's thesis at Technical University of Munich (TUM) involved a neural network called Hierarchical Temporal Memory (HTM). I used and adapted Numenta's technology NuPIC. HTMs, as claimed by Numenta, are inspired by the human brain. I achieved 1-second prediction of motion for simple tasks, but the network was unable to generalize to more complex tasks. With help by my former supervisors, I turned the results of my Bachelor's Thesis into a paper, which I presented at IEEE International Conference on Cyborg and Bionic Systems (CBS).

I created music management software. The software is free and open source, and offers many features not found in iTunes or other software, such as music analysis, a novel frequency based visualizer, and integration into DJ software. The earlier project, now mature, is targeted at DJs. The later project is still in development, but is music source agnostic (e.g. with support for spotify).

For one of the modifications I made for the game Minecraft, I created a filehosting website. On this site, people could share the creations they had made. It has account and content management systems, as well as a popularity index.

I worked at the department for education at the Technical University of Munich (TUM). My main tasks involved setting up, maintaining, and expanding the educational software Open edX.

I created a client for a file hosting server. The client supported quick screenshot, file, and text snippet uploads to share with other people. The app supported two server backends, both of which shut down by now.

I created the first version of my personal website in Laravel. I have since dropped this version in favour of a leaner, smaller tech stack version.

During a practicum at Xyrality, I contributed bug fixes to their game Lords and Knights.

I created several modifications for the game Minecraft during my late high school and early university life. These mods achieved a substantial popularity, standing at more than 100,000,000 total downloads today. These projects involved a lot of community interaction, and active maintenance under a rapidly changing platform. Most projects were solo, though a few involved collaboration with other creators.

During a practicum at a mobile software company (lb-lab), I designed and created a mobile app intended for claiming damages in case of commute accidents.

I occasionally create small video games. This is an ongoing habit since my early school years, and I will likely continue doing this to hone various skills or test out ideas. I made a cozy interactive experience, an arcade bullet hell, an idle clicker, a digital board game, and a 2D survival horror game.

I started programming with several smaller GUI projects. None of these are available anymore, but they included a website for a business, a clock app for iPhone, and a macOS app to catalogue video games.