Meet the students: Sara Abdelrazeq

From Jordan to Geneva, Sara K. Abdelrazeq has joined CERN this summer through the CERN openlab programme, contributing her skills in computer science to the Next Generation Triggers project. Currently a student at the University of Bonn, she is working on one of the most critical aspects of modern particle physics: the real-time systems that decide which collision events to record for analysis.

With the upcoming upgrade of the CMS trigger system, machine learning will play an increasingly central role in processing vast amounts of detector data. But because detector conditions evolve over time, algorithms must be able to adapt, without forgetting what they have already learned.

Sara’s project explores how to make this possible. She is investigating continual learning methods for decision tree-based algorithms, developing strategies to update triggers incrementally as conditions change, and testing training approaches that preserve performance while enabling adaptability. Her work also looks at how these methods can be implemented directly in the CMS trigger ML training infrastructure.

By advancing continual learning for CMS, Sara’s research is helping to ensure that the experiment’s trigger system remains robust and flexible in the high-luminosity era — ready to capture the next big discovery.