Selected work
Benchmarking time-series foundation models for gyrokinetic plasma heat-flux prediction
MSc thesis in progress, expected 2026. Johannes Kepler University Linz, supervised by Fabian Paischer.
Turbulent heat flux in a tokamak is a single scalar, but obtaining it normally means integrating the 5D distribution function of a nonlinear gyrokinetic simulation, which is expensive. I am testing how close fine-tuning and in-context learning on time-series foundation models, given the scalar flux time series alone, can get to the accuracy of the GyroSwin-1B 5D neural surrogate, without modelling the 5D distribution function the way GyroSwin does.
Benchmarked against Paischer et al., GyroSwin: 5D Surrogates for Gyrokinetic Plasma Turbulence Simulations, NeurIPS 2025. arXiv:2510.07314
Apollonia, handwriting recognition for dental records
Contract project for AKS-ZAVOMED, an Austrian NGO, 2025 to 2026.
The project covers around 50,000 scanned handwritten dental examination forms. I trained a custom CNN, roughly 120,000 to 150,000 parameters over 12 classes on 32x32 grayscale regions of interest, to read per-tooth codes, reaching 96.93% per-cell accuracy on 3,257 of 3,360 cells. That took an automated experiment loop over 19 or more variants, plus around 1,200 hand-labelled cells to close the gap between public handwriting datasets (EMNIST, ARDIS, DIDA) and Austrian digit conventions. It runs fully on device to satisfy GDPR: ONNX INT8 export for CPU inference on an Apple M1 Max, with a PySide6 review interface for the operator.
Brainchair, open-source MCP server
Go, 2026.
An always-on Model Context Protocol server that gives Claude Code, Claude Desktop and other clients one shared SQLite memory, with FTS5 full-text search, filters on kind, domain, project and tags, and version history kept by database triggers. Entry types are registered at runtime and generate their own SQL views, indexes and typed MCP tools with JSON schemas. In daily use as my own knowledge store and CRM, currently 265 entries across 8 types.
github.com/lukaskurz/brainchair
Vineyard row detection from satellite imagery
BSc thesis, completed 2025. Johannes Kepler University Linz, supervised by Jan Schlüter. Grade sehr gut (1).
I annotated the dataset myself, 1,422 rows across 79 Austrian sites in QGIS, from 30 cm basemap.at orthophotos and INVEKOS parcel boundaries. A U-Net trained on four-channel patches (RGB plus parcel mask) reached 0.99 ± 0.12 m spatial accuracy and 88% row-count accuracy, inside the tolerance agricultural GPS needs. I compared loss functions for thin linear features and defined a row-count metric, because overlap metrics punish one-pixel offsets that do not matter in the field. The best models used centerline Dice, a topology-preserving loss, over the standard Dice that earlier experiments had favoured.
Read the thesis (PDF)
Vinitor
Founder and sole engineer, Vinitor Software Solutions e.U.
Since September 2025.
Vinitor does GPS time tracking, plant-protection documentation, and plot and equipment management for vineyards and agricultural operations in the DACH region. I build and run the entire stack alone: a Kotlin/Spring Boot backend, PostgreSQL/PostGIS, an Angular dashboard, a Flutter app in 16 languages, GitLab CI, Hetzner and Cloudflare infrastructure, and Stripe billing. It integrates the national geodata services each country publishes, among them the AMA/INVEKOS OGC API, basemap.at, SWISSIMAGE, BKG DOP20 and the Czech LPIS. Everything customer-facing is mine too, from sales and onboarding to support, releases and the newsletter, in German and English. Development runs on agentic workflows, which is what keeps a product this size maintainable by one person.
Co-founder, then technical managing director, Vinitor GmbH
September 2021 to 2025.
I joined during Covid as the engineer needed to turn a colleague's diploma project into a product, and moved into the technical lead role as the team grew. I owned the complete technical side of an eight-person company, hired and led three software engineers and a UI/UX designer, and built the Flutter app, the CI/CD pipelines and the server infrastructure while working across the Angular dashboard and the Spring backend. The company reached 25 to 35 customers at its peak and entered insolvency in early 2025. I let the technical team go, took over the intellectual property, and have continued the product alone at my own cost since.
Skills
Python, PyTorch, ONNX, Kotlin, Go, TypeScript; neural surrogates, time-series models, CNNs, on-device inference, LLM agent and RAG workflows.
