Max Robotics — Global Service Robot Directory + US Certification

About

Alexander Khazatsky is a leading researcher in robotics, with a focus on large-scale robot learning, manipulation, and self-supervised skill acquisition. His most impactful contributions center on the creation of massive, diverse datasets and models that enable general-purpose robotic policies. He is a key contributor to the **Open X-Embodiment** collaboration, which produced the RT-X models—high-capacity models trained on a vast collection of robot data from over 20 institutions. This work, with over 220 combined citations, has been pivotal in demonstrating that, like in NLP and computer vision, scaling data and model size can unlock robust, cross-embodiment robotic capabilities. Khazatsky also leads the **DROID** project, which has generated one of the largest in-the-wild robot manipulation datasets (over 100 citations), addressing the logistical challenges of collecting diverse, real-world data. Earlier, his work on **Contextual Imagined Goals** and **DisCo RL** advanced self-supervised and distribution-conditioned reinforcement learning, enabling robots to propose and practice their own goals for more flexible skill acquisition. Through these efforts, Khazatsky is helping to consolidate the field toward generalist robotic policies.