Azra Begzadić

Email: abegzadic (at) ucsd.edu

Azra Begzadić is a Ph.D. student in Mechanical and Aerospace Engineering at UCSD. Her research interests lie in safety-critical learning-based control.

Before joining UCSD, Azra received an M.Sc. from the Technical University in Munich, Germany, in 2023, and a B.Sc. from the Vienna University of Technology, Austria, in 2020 (both in electrical engineering and information technology).

In her free time, Azra enjoys photography, drawing, and reading books.

Selected Projects

Event-triggered learning for safe control: This framework uses event-triggered learning, which switches between prioritizing control performance and improving model accuracy based on the uncertainty of the learned model.

Safe reinforcement learning with reach-avoid safety filters: Using reach-avoid sets to minimally override the RL process while ensuring the system (a) maintains safety and (b) returns to its base configuration at the end of the horizon.

Related Papers

  • Azra Begzadić, Armin Lederer, Jorge Cortés, and Sylvia Herbert. “Learning high-order CBFs using Gaussian processes for systems in Brunovský canonical form.” Proceedings of the IEEE Conference on Decision and Control, Rio de Janeiro, Brazil, 2025, to appear.

  • Sander Tonkens*, Nikhil Uday Shinde*, Azra Begzadić*, Michael C Yip, Jorge Cortés, and Sylvia Herbert. “Space to Time: Out-of-Distribution Generalization of Safety Filters via Temporal Disturbance Encoding.” Second Workshop on Out-of-Distribution Generalization in Robotics at RSS, 2025.

*co-first authors