Precision Flavor Measurements and Real-Time Anomaly Detection at the CMS Detector

As a part of the CMS collaboration, I have been actively involved in cutting-edge projects that explore precision measurement in particle physics and advanced machine learning techniques, with a focus on testing lepton flavor universality (LFU) and developing real-time anomaly detection systems. One of my key endeavors is the precision measurement of the R(K) ratio, a critical test of LFU and the Standard Model. Using novel data-taking techniques like "B Parking" event storage and dynamic trigger thresholds, I am working towards producing the most precise R(K) measurement done with a general-purpose detector to date. In parallel, I am working on AXOL1TL, an unsupervised anomaly detection algorithm. This algorithm uses an autoencoder to select interesting events without bias toward specific physics signatures. Implemented on FPGA hardware at the CMS Level-1 Trigger, AXOL1TL reads in detector inputs and makes decisions in nanoseconds. My contributions include testing network architectures for timing and precision, integrating the algorithm into the existing trigger firmware, and developing a monitoring framework to ensure stability. Following its successful commissioning in May 2024, I continue to analyze initial data and support the first analyses using this innovative trigger.