Backend systems
I work on web applications people actually use: APIs, internal tools, data models, and robust systems.
AI researcher · backend developer · Linux administrator
I build software, research AI, and look after Linux systems. Most of my work is somewhere around backend development, applied AI, education, and automation.
What I do?
I work on web applications people actually use: APIs, internal tools, data models, and robust systems.
I like the point where models meet messy data: LLM evaluation, scanned tests, handwriting, video, and tools that make experiments easier to repeat.
I run Linux infrastructure with tools I know I can debug: Ansible, Docker, Kubernetes, monitoring, and deployment scripts that do not try to be clever.
Research
I study how language models solve academic problems, and build systems that use AI on data from the real world: student answers, scientific workflows, and video.
Publications
EMNLP 2025. We built a benchmark from Slovak math, physics, and informatics competitions. The goal was simple: check whether language models can solve hard competition problems.
Current work
I work on AI for video. The useful version is easy to describe: find what is happening, when it happens, and which moments are worth keeping. The technical parts include action spotting, live video processing, temporal segmentation, and broadcast tools.
Research direction
I am interested in what happens when several AI assistants work together instead of one model doing everything alone. Can smaller local models cooperate well? When should a teacher or researcher stay in control? Where can LLMs help with feedback, grading, or scientific work?
Teaching & education
At FMFI, I teach courses where students turn code into working systems. I also spent several years in programming education outside school, mostly through KSP at Trojsten and coding activities for high school students.
Winter
Sorting, priority queues, dictionaries, text search, greedy algorithms, dynamic programming, divide and conquer, graph algorithms, and the basics of computational complexity. I lead tutorials and helped refresh course materials.
Summer
Object oriented programming in Java: classes, interfaces, collections, I/O, serialization, concurrency, JavaFX, and applications with a user interface. I work on grading and exams.
A practical course on Docker, container images, networking, Docker Compose, Kubernetes basics, Helm, monitoring, and CI/CD deployment pipelines. I lead the subject and lectures.
Contact
[email protected]