Max Robotics — Global Service Robot Directory + US Certification

About

Kirsty Ellis is a robotics researcher whose work spans 3D scene understanding, robot manipulation, and mobile robot navigation, with a particular focus on enabling robots to operate intelligently in complex, real-world environments. She has made notable contributions to open-vocabulary scene representation through her involvement in **ConceptGraphs**, a framework that leverages large vision-language models to construct semantically rich 3D scene graphs for robot perception and planning — a paper that has garnered an impressive 178 citations since its 2024 publication. Ellis also contributed to **DROID**, a large-scale in-the-wild robot manipulation dataset (108 citations) that addresses one of robotics' most pressing challenges: generating diverse, high-quality training data at scale. Her earlier work on garbage collection and sorting using deep learning and whole-body control (34 citations) demonstrated practical real-world applications of intelligent mobile manipulation. A recurring theme across her research is **Navigation Among Movable Obstacles (NAMO)**, where she has advanced the field by developing systems that allow robots to actively interact with and reorganize their environments rather than merely avoiding obstacles. Collectively, Ellis's research reflects a commitment to bridging perception, learning, and physical interaction in robotics.