Tianhong Li is a PhD candidate whose research utilizes machine learning (ML) approaches in the development of novel sensing systems. His research has provided important new insights into ML problems, such as learning from unbalanced datasets and improving the robustness of self- supervised learning, enabling it to deal with new modalities. A MathWorks Fellowship will support Tianhong’s work on several exciting projects. The first is a groundbreaking system that provides highly accurate human pose estimation through walls and occlusions by leveraging the properties of RF signals in Wi-Fi frequencies. The second is RF-Pose3D, the first system that infers 3D human skeletons from RF signals; potential applications include gaming, surveillance, and healthcare. Finally, he is conducting a project in unsupervised learning for signals beyond human perception, to leverage unsupervised learning algorithms to utilize unlabeled RF data for pre-training. Tianhong’s algorithms are already in use to monitor the motion of patients with Parkinson’s disease and other motion disorders, and his wide-ranging work offers tremendous promise to advance unique sensing and ML applications and bring new capacity to MathWorks-driven research.