I am a programmer and Computer Science researcher, strongly rooted in the open source hacker culture. My focus is Artificial Intelligence and Linux system programming. As another hobby, I enjoy messing with digital circuits and I like the game of Go. I also like Doing Things Right™.
I'm the co-founder of Rossum, a London / Prague startup providing a leading solution for Cognitive Data Capture from Documents using bleeding edge computer vision AI.
Before, we co-founded Ailao, a Prague AI company working on user-friendly machine learning on unstructured data like images and documents.
My current main research focus is
Information Extraction (from natural language data) and Machine Learning.
I have created the YodaQA
question answering system (think IBM Watson DeepQA)
and the dataset-sts
framework for processing sentences with neural networks.
I am also the main author of the world-class Go playing
program Pachi,
and did some work in continuous black-box optimization.
I have contributed to my fair share of open source projects; most notably, I have been working on Git, ELinks and glibc. Almost all my personal pet projects are open source too. I have also conjured up a couple of web services, did not shy away from some reverse engineering and done a wide variety of consulting work.
For quite a few years, I have been teaching a variety of courses (mostly related to UNIX, high performance computing and AI) at the Charles University; I also taught at the Czech Technical University.
I am quite involved with the Prague hackerspace. I work on most of my hobby projects there, mostly involving some hardware and microcontrollers. Doors, robots, scientific experiment apparata, sensors, complex LED control, you name it!
That said, building Rossum is taking most of my energy and time nowadays.