Using data from three recently completed pragmatic clinical trials at the Yale School of Medicine (the ELAIA 1, ELAIA 2, and KAT trials), which evaluated electronic health record-based alert interventions for kidney care, this project focuses on exploring treatment effect heterogeneity with machine learning and discovering meaningful subgroups. Students will be organized into small working groups, each focusing on one trial. They will apply state-of-the-art causal machine learning methods, such as the Meta Learner, Bayesian Additive Regression Trees, and Bayesian Causal Forest, to explore individualized treatment effects and assess whether certain patient subpopulations benefit or do not benefit from alert-based healthcare interventions. Students will critically evaluate the assumptions behind different subgroup identification approaches and compare findings across methods. They will also collaborate across working groups to understand how trial settings affect appropriate methods and results. This is an open-ended pursuit designed to encourage innovative thinking, teamwork, and hands-on implementation, simulating real-world team science in clinical research. The project offers rigorous training in modern causal inference techniques for heterogeneous treatment effects and their application to clinical trial data.