Hi, I'm Janos.

A lifelong passion

From the moment I first opened a computer as a child, I've been fascinated by how things work. That early curiosity grew into a lifelong love for IT, driving me to constantly challenge my mind by working out solutions to tough problems. Whether tinkering with hardware, writing low-level code, or unraveling security puzzles, I thrive on the intellectual thrill of discovery and innovation.

What drives me

One driver has stayed constant across every project: connecting the software world to the real world. We are living through the moment where decades of pure-code work is finally meeting physical systems that sense, act, and decide for themselves. That is where my interest has always lived, and the thread that ties everything below together.

• Autonomous behaviour: the moment a robot stops following a script and starts making its own decisions. From perception to recovery, the goal is machines that cope with the real world without a human in the loop.
• Pushing the boundaries of AI: applied experimentation where AI meets the physical world. Not benchmarks or clean demos, but systems that act, fail, and learn in environments that do not cooperate.
• A lifelong love of Linux: twenty-plus years in the Linux ecosystem, from userland to kernel. It is the substrate every interesting system I have built sits on, and the one place I will always want to be hands-on.

Current focus

Today, most of my hands-on time goes into humanoid robotics: teleoperation and leader-follower control on the Unitree G1 and AGIBOT A2, autonomous behaviour development, and the practical question of where humanoid platforms actually earn their keep. In parallel, quadruped autonomy on the Unitree Go2 for industrial inspection and construction surveying, and the multi-robot collaboration work that has run through my research for years.

Where robotics is going

We stand at the brink of a technological revolution in robotics. Today's hardware can achieve smooth, lifelike motion once reserved for science-fiction, but true progress lies beyond mechanical precision.

Movement alone does not create a robot worthy of the title "humanoid." The breakthrough will come not from the perfection of motors, joints, or sensors, but from the emergence of autonomous, simulated intelligence. Only when robots can think, adapt, and make decisions independently, navigating the unpredictable nature of the real world, will they truly cross the threshold from machine to autonomous entity.

My focus:

• Advanced autonomy: simulated intelligence that lets robots think, adapt, and decide independently in unpredictable environments.
• Multi-robot cooperation: collaborative behaviours among fleets of robots to unlock new levels of efficiency and capability.
• AI-powered physical-world integration: fusing cutting-edge mechanics with self-directed intelligence, so machines are not just humanoid in form, but autonomous in function.

This is the real frontier: the fusion of advanced mechanics with self-directed, intelligent behaviour. It's here, in the realm of autonomy, that the next great leap in robotics will be forged. And when it happens, it won't just redefine what a robot is, but what it means to coexist with them.