Harald Reingruber

Led the technical transformation of Dedalus’ medical 3D visualization module, migrating from a C++/DirectX 9 codebase to a modern Rust+WebAssembly+WebGL/OpenGL and TypeScript framework. Conducted comprehensive technology evaluations to select the optimal stack for web deployment. Repackaged the module as a framework-agnostic Web Component, now shipping in production in two host products (Angular- and React-based). Engineered a robust side-by-side comparison system, facilitating precise performance and visual validation during the transition. Established a component-level E2E testing layer with Storybook and Playwright, replacing cumbersome end-to-end tests in host applications and ensuring software reliability. This modernization initiative enhanced performance, expanded platform reach, and positioned Dedalus at the forefront of medical visualization technology.

The three10 software allows delivering presentations including 2D and 3D media content synchronized across multiple devices. For example, a mobile device can be used for remote-controlling the presentation while a high-end graphics device is visualizing the presentation on a high-resolution video wall.

As one of two principal engineers in a team of 6 developers, I am the main responsible engineer for the secure communication and file transfer between the Unity 3D application and the on-premise backend server. The architecture of the TypeScript/React/Node.js backend software is also one of my responsibilities.

Most notable achievements:

• Implemented a JavaScript library for streaming server-side Unity 3D graphics in real-time from a cloud server to an HTML5 browser using a low-latency video codec.
• Implemented HTML5 Server-Sent-Events for pushing data updates from the backend server to the Unity 3D client software (C# client implementation, as well as Node.js server implementation).
• Implemented a Docker deployment pipeline for our on-premise backend software, in order to minimize the impact of running the software on various operating systems and versions.
• Implemented screen capturing and streaming over the network.
• Became the go-to guy for Git, Docker, Gitlab, Gitlab CI and Unity Cloud Build.
• Principal technical contact during the Audi software security certification process.

See the three10 multi-client presentation and collaboration software in action:

Intestinal Content Detection in Capsule Endoscopy using Robust Features

The aim of this project was to develop a method supporting the automatic analysis of capsule endoscopy studies in order to handle the large amount of image data in the daily clinical routine.

Most notable achievement:

• Implemented a Support Vector Machines (SVM) detector for detecting intestinal content (71 % accuracy)

In a team of 22 people, I was responsible for implementing and maintaining 3D visualization features and the loading of volume image data.

Most notable achievements:

• Implemented progressive loading of volume image data (patent claimed)
• Visualization of the normal plane in the unfolded blood vessel view
• Improved visualization of the intersection of 2D annotations with the current view plane (semi-transparent rendering of the part overlapping with the adjacent slice)