Hi!

I am Matej Grcić, PhD candidate at University of Zagreb, Faculty of Electrical Engineering and Computing. Under supervision of prof. Siniša Šegvić, my research focuses on enhancing the safey of deep learning models deployed in the real world. Specifically, I work on open-set and outlier-aware semantic segmentation with a strong emphasis on real-time applications such as autonomous driving. I also worked on semantic segmentation of traffic scenes with adverse driving conditions, which resulted in winning CVPR’22 Vision4AllSeasons competition.

In parallel, I am also a visiting PhD at EPFL, supervised by prof. Maria Brbic and funded by Swiss Government Excellence Scholarship. Here, my research explores fundamental machine learning topics, including recognition and representation learning under limited supervision. This research has led to the development of weakly-supervised models capable of discovering fine-grained subclasses.

Prior to my PhD studies, I earned both my Master’s and Bachelor’s degrees from the University of Zagreb, where I was honored with the Rector’s Award for outstanding individual research. During my studies, I also gained valuable industry experience through internships at LifeNome, Axilis, and CROZ.

I am always open to discussions and collaborations. Feel free to reach out to me via email.