A combined TCAD and machine learning workflow offers a way to narrow the search for bridging defects in FinFETs. By learning from simulated electrical characteristics, a Random Forest model predicted defect regions with high average accuracy in the study.