Max Robotics — Global Service Robot Directory + US Certification

About

Marianna Madry is a robotics researcher whose work sits at the intersection of tactile sensing, robotic grasping, and machine learning for embodied intelligence. Her research focuses on enabling robots to perceive, learn, and generalize manipulation strategies across diverse objects — a fundamental challenge in making robots practically useful in unstructured environments. Madry's most influential contribution is ST-HMP (Spatio-Temporal Hierarchical Matching Pursuit), an unsupervised feature learning framework for tactile sensor data, which has garnered 116 citations and represents a significant advance in how robots can interpret rich tactile information during grasping and object recognition tasks. Complementing this, her work on learning dictionaries of prototypical grasp-predicting parts (100 citations) demonstrated how robots can transfer grasping strategies across novel objects by identifying shared structural features — bridging the gap between example-based learning and generalizable robotic behavior. Her research on grasp generalization and task-specific object representation further underscores a consistent theme: equipping robots with human-friendly, transferable knowledge rather than object-specific programming. With contributions to humanoid robot grasp adaptation and object categorization, Madry has helped lay important groundwork for intelligent, adaptive robotic manipulation systems that learn meaningfully from experience.