About

Hi, I’m Lazar! I’m a researcher working on computer vision and generative AI, with a particular interest in how models represent visual information and how we can make them more efficient and robust.

I finished my undergrad in Computer and Information Science at the University of Ljubljana, where I received the Dean’s Award for Outstanding Academic Achievement and worked with Prof. Marko Robnik-Šikonja on sarcasm detection in Slovene. I recently completed the Erasmus Mundus Joint Master in Artificial Intelligence, studying at Pompeu Fabra University, Radboud University, and the University of Ljubljana on a full Erasmus Mundus scholarship and graduating cum laude. For my master’s thesis, I worked with Prof. Matej Kristan and Dr. Alan Lukežič on memory-based methods for long-term visual point tracking.

Currently, I’m a research intern at EPFL’s VILAB, working with Prof. Amir Zamir on efficient visual tokenization for generative models. Previously, I interned at Microsoft and Teads, working on AI agents for understanding code and efficient training of prediction models.

I have broad interests and always enjoy a good discussion. If you’re working on something interesting or want to exchange ideas, feel free to reach out.