Max Robotics — Global Service Robot Directory + US Certification

About

Oier Mees is a robotics researcher whose work sits at the intersection of language grounding, robot learning, and embodied AI. His research focuses on enabling robots to understand and execute complex tasks specified through natural language, particularly by leveraging large-scale pretrained vision-language models and imitation learning from unstructured data. Mees has made significant contributions to language-conditioned robot navigation and manipulation. His work on Visual Language Maps (301 citations) demonstrated how off-the-shelf vision-language models can ground natural language instructions to a robot's visual observations for goal-directed navigation. Complementing this, his research on language-conditioned imitation learning (68 citations) and visual affordance grounding (66 citations) has advanced how robots learn generalizable multi-task skills from unstructured demonstrations with minimal human intervention. He also contributed to the influential Octo generalist robot policy (66 citations), reflecting his engagement with scalable, open-source robot learning frameworks. Earlier in his career, Mees explored spatial relation generalization and self-supervised 3D shape estimation for robotics. More recently, his work on efficient action tokenization for vision-language-action models (28 citations) addresses critical challenges in deploying transformer-based robot policies. Across his career, Mees has consistently pushed toward robots that perceive, reason, and act in human-centered environments with greater autonomy and generalization.