About me

Page

I am a Software Development Engineer at Skype (Microsoft). I work on SMS backend, taking care of end to end delivery of SMS.

I finished a Master’s Degree in Theoretical Computer Science at Faculty of Mathematics and Physics, Charles University in Prague in 2013. My research topic was Approximative Bayes methods for belief monitoring in spoken dialogue systems, my thesis advisor was Filip Jurčíček.

I used to organize Czech correspondence seminar in programming. I am still interested in computer science education.

I spent some time working with NLP Research team at Google Switzerland, where I worked on opensourcing TopicSum summarizer and on some other cool projects.

Contact

E-mail / Google talk / xmpp: david@marek.me
LinkedIn: http://www.linkedin.com/in/davidmarek

Projects

Vystadial

The goal of the project is to study and improve statistical methods for learning of statistical models used in complex dialogue systems. My task is to develop approximative methods for belief state update. The project is lead by Filip Jurčíček and we work on it together with Matěj Korvas, Ondřej Dušek, Ondřej Plátek and Lukáš Žilka. We have developed a dialog system for getting public transport directions in Prague, called Alex on the Bus.

Prak

Prak stands for Programmers Academy, this project was created to promote computer science among Czech high school students. Our goal is to create an online framework for programming courses similar to codecademy, udacity and other education startups. I work on this project with my friends Lukáš Lánský, Jiří Maršík and Tomáš Jakl. We are supported by Faculty of Mathematics and Physics.

Software

E-Puck library

Python library for controlling e-puck robot remotely from a PC. Developed as a part of my Bachelor’s thesis. It is used in Evolutionary Robotics course.

TopicSum

Opensourced TopicSum summarizer [Haghighi, 2009] and modified to use data from Enrycher. This summarizer is used for summarizing news articles, it uses topic model for selecting best sentences from an article to add to summary. Used in RENDER project in collaboration with Institut “Jožef Stefan” for diversified news search engine.

Publications

Approximative Bayes methods for belief monitoring in spoken dialogue systems, Master thesis, Faculty of Mathematics and Physics, Charles University in Prague, 2013

Controlling e-puck robot in Python, Bachelor thesis, Faculty of Mathematics and Physics, Charles University in Prague, 2011.

Lukáš Žilka, David Marek, Matěj Korvas, Filip Jurčíček. 2013. Comparison of Bayesian Discriminative and Generative Models for Dialogue State Tracking.
In 14th Annual Meeting of the Special Interest Group on Discourse and Dialogue (SIGDIAL).