In this session, I will use multimedia videos and research findings to lead a discussion about student learning during online problem-solving practice in mathematics and science. We will discuss characteristics that sometimes prevent students from gaining deep, long-lasting understanding of math and science concepts, especially when working with computerized practice tools. I will demonstrate the features of a computer-based Intelligent Tutoring System that has been developed for geometry, and discuss how we’ve used this system to investigate student learning and identify problematic student behaviors and thinking during computer-supported mathematics practice. Finally, I will present practical implications from the research findings, both for geometry as well as other math and science domains. We will discuss how teachers can support deep student learning, with and without computer technologies.