Max Robotics — Global Service Robot Directory + US Certification

About

Katerina Fragkiadaki is a leading researcher at the intersection of robot learning, computer vision, and natural language understanding, with a particular focus on enabling robots to perceive, reason, and act in complex real-world environments. Her work spans embodied AI, visuomotor control, generative simulation, and 3D scene understanding, reflecting a broad and deeply integrated research vision. Fragkiadaki's most influential contributions include pioneering frameworks that leverage large language models for open-ended robot instruction-following, energy-based compositional planning for scene rearrangement, and generative simulation pipelines that dramatically reduce the human effort required to train robotic manipulators. Her RoboGen and Gen2Sim systems exemplify her commitment to scalable, automated robot learning, while Act3D demonstrates her expertise in spatially grounded manipulation using 3D feature representations. Earlier work on reward learning from narrated demonstrations and graph-structured visual imitation established her as an innovator in connecting natural language and visual perception to robotic behavior. With papers accumulating citations across robotics, vision, and AI communities, Fragkiadaki's research has meaningfully shaped how modern robotic systems are trained, instructed, and evaluated. Her work is especially valuable for students and researchers seeking to understand how foundation models, 3D perception, and generative tools can converge to create more capable, generalizable robotic agents.