Senior Research Scientist - Robotics Foundation Models

Maxinsights•Santa Clara, CA

About The Position

Maxinsights is building the foundation layer for Physical AI, the intelligence that enables robots to perceive, reason, learn, and act in the real world. Our mission is to unlock scalable learning for robotics through data and foundation models trained on real-world interaction. We believe the next generation of robotics will be powered by such models, and we're building the data and intelligence layer to make that possible. We are looking for a world-class Senior Research Scientist to define and lead our research agenda in Robotics Foundation Models. This is a highly research-oriented role with significant scientific autonomy and long-term impact. You will shape the direction of embodied AI at Maxinsights, lead research into large-scale multimodal models, and work at the intersection of data, model architecture, and real-world robot interaction. You will have the opportunity to pursue ambitious research questions, build new approaches to learning from embodied data, and turn fundamental advances into systems that can operate across a wide range of robotic platforms and tasks.

Requirements

PhD in Robotics, Machine Learning, Computer Vision, Computer Science, or a related field.
Deep research expertise in robotics foundation models, embodied AI, multimodal learning, or a closely related area.
Hands-on experience designing and training large-scale multimodal or foundation models.
Strong understanding of modern deep learning architectures, representation learning, and model scaling.
Strong publication record in leading machine learning, computer vision, or robotics conferences.
Demonstrated ability to independently identify important research problems and drive them from idea to execution.
Strong communication and collaboration skills, with the ability to work effectively across research and engineering teams.

Nice To Haves

Experience at a leading AI or robotics research organization such as OpenAI, Google DeepMind, Meta FAIR, Physical Intelligence, or similar.
Hands-on experience with VLA models, robotics pretraining, world models, or large-scale policy learning.
Experience scaling distributed training using technologies such as FSDP, DeepSpeed, Megatron, TPU, or equivalent systems.
Experience working with large-scale real-world robotics or egocentric datasets.
A strong point of view on representation learning, generalization, scaling, and the path toward general-purpose robotic intelligence.
Track record of turning fundamental research into systems that work in real-world environments.

Responsibilities

Define and drive the long-term research agenda for Robotics Foundation Models at Maxinsights.
Lead research and large-scale training of multimodal models spanning vision, language, and action.
Shape the design, composition, and scaling strategy of our embodied datasets.
Develop novel approaches to learning representations and policies from large-scale real-world interaction data.
Investigate how foundation models can generalize across robots, environments, tasks, and embodiments.
Translate research ideas into scalable systems in close collaboration with engineering and robotics teams.
Publish high-impact research at leading venues such as NeurIPS, ICML, ICLR, CVPR, RSS, CoRL, and ICRA.
Mentor researchers and help establish the scientific culture and research direction of a world-class robotics organization.