5th year student as Master of Science in Media Technology and Engineering
/ status
High-performance GPU-driven voxel terrain engine in C++20 with OpenGL 4.3+ compute shaders, indirect drawing, spline-based terrain shaping, and a multi-threaded chunking architecture.
Autonomous drone navigating procedurally generated terrain in C++/OpenGL — control loop steers the drone through the simulated environment in real time.
Real-time gesture recognition driving a Godot game over WebSocket — MediaPipe hand tracking, a TensorFlow/Keras CNN classifier, and an OpenCV webcam pipeline.
GPU path tracer in C++/OpenGL with fragment-shader rendering for photoreal frames.
Unity 3D 1v1 RTS using ArUco cards for unit placement; built for my bachelor thesis.
Blender add-on to simulate and auto-animate solar systems with generated materials.
Research papers and writeups.
Jonatan Ebenholm
Compares two approaches to real-time gesture recognition for game input: a custom CNN trained on the HaGRID dataset plus self-collected images, versus a Feed Forward Neural Network operating on 21 3D hand landmarks extracted with MediaPipe. The MediaPipe + FFNN hybrid reached 100% validation accuracy with sub-second per-epoch training, outperforming the CNN on both accuracy and rotation robustness.
Ludwig Boge, Jonatan Ebenholm
A path tracer implemented entirely in a fragment shader (C++/OpenGL), targeting near real-time rendering of diffuse, specular, glossy, and transmissive materials. Accelerated with a CPU-built Bounding Volume Hierarchy traversed in the shader via Shader Storage Buffer Objects.
Ludwig Boge, Jonatan Ebenholm
A Monte Carlo raytracer written from scratch in C++ for the TNCG15 Advanced Global Illumination course. Solves the rendering equation via Monte Carlo estimation across Lambertian reflectors, perfect mirrors, transparent surfaces, and area light sources; benchmarks the effect of shadow ray count and sample count on render quality.
Berkay Orhan, Jonatan Ebenholm
A content-based recommendation system that suggests anime titles from intrinsic metadata using BERT embeddings and cosine similarity. Avoids the cold-start and privacy concerns of collaborative filtering — built around a quick "use and leave" web flow with no user accounts.
SICK IVP / Linköping University
Master's thesis applying inverse rendering — recovering scene geometry, materials, and lighting from images — to industrial inspection workflows.
Linköping University
Paid programming assistance role under LiU's LiTHeHack initiative, supporting students with programming work.
Linköping University
Lab assistant for the Applied Transform Theory course — guided students through lab sessions on Fourier analysis, Laplace transforms, and signal/system analysis.
Linköping University
Specializing in Computer Graphics, GPU programming, video game systems programming, and machine learning / AI.
Linköping University
Foundations in programming, applied mathematics, and media technology.
Selected highlights from my MSc program.
I'm finishing an MSc in Media Technology and Engineering at Linköping University, with the focus pulled hard toward computer graphics, GPU programming, and applied machine learning. My master's thesis is hosted at SICK IVP in Linköping, where I'm working on inverse rendering for industrial inspection — recovering scene geometry and materials from imagery to support automated quality assessment.
I came at graphics through art — years of figure and gesture studies, then a media technology degree for the practicality plus the graphics angle I'd been curious about as a gamer. Code turned out to be a way to build worlds the way drawing built faces and figures. TNCG15 — Advanced Global Illumination is what made me a graphics programmer proper, and TSBK07 — Computer Graphicsthe semester after dropped me into parallel programming. I haven't really left.
Most of what I build lives on the line between graphics and systems: GPU-driven voxel engines, fragment-shader path tracers, procedurally generated worlds. I like the work that asks you to think at two levels at once — the high-level structure of what's being rendered, and the low-level mechanics of how the GPU is actually executing it. The projects above are mostly me chasing that.
After graduation, I'd love to land on a graphics or rendering team at a game studio, or industry-adjacent work in the same space as my thesis — vision, simulation, perception. A full-time research role would tempt me too, if the project is the right kind of strange.
