Applied control-theoretic tools to analyze deep sequence models (state space models, transformers);
designed and experimentally validated uncertainty-aware, safe perception-based control for autonomous racing;
developed a SLAM method with explicit uncertainty guarantees for robust perception in safety-critical settings.
Coding instructor
Girls Code Too
Instructed Python coding classes, with a special focus on preparation for the Swiss Informatics Olympiad.
Research in Industrial Projects for Students
Institute for Pure and Applied Mathematics, University of California Los Angeles
Researched automatic conflict detection in police body-worn camera audio, developing a novel method combining adaptive noise removal, ML-based speech segmentation, and conflict measures from phrase repetition and intensity
Education
PhD (System control and computer vision focus)
ETH Zurich
Thesis focuses on the interface between control and vision, enabling safe perception-aware control of autonomous systems.
MS Electrical Engineering (Machine learning and signal processing focus)
ETH Zurich
GPA: 5.8/6.0 (with distinction)
BS Electrical Engineering and Computing
School of Electrical Engineering, University of Belgrade
GPA: 10.0/10.0 (student valedictorian)
Focused on system control, signal processing and machine learning.
Skills & Hobbies
Technical Skills
Python, PyTorch
C++, MATLAB
JAX, TensorFlow, ROS
Hobbies
Photography
Learning languages
Reading
Awards
Excellence Scholarship and Opportunity Programme Scholar 2018