About the role Overview RealHand is an AI robotics company building dexterous robotic hands and the intelligence that drives them. The goal of the company is to give machines human-level dexterity — hardware that survives real-world contact, and policies learned from how humans actually use their hands. Our hands are deployed on humanoid and mobile platforms across research, automation, and human-robot collaboration. RealHand is headquartered in Palo Alto, CA. We are looking for full-time Research Engineers and Research Interns to develop algorithms and systems for embodied AI and robotic manipulation. The work focuses on enabling robots to understand multimodal instructions, learn from data and physical interaction, and perform complex manipulation tasks using robotic arms, dexterous hands, and bimanual systems. Research areas include foundation models, robot learning, perception, manipulation, planning, reasoning, and real-world robot deployment. This is an algorithm-focused role, with physical robotic systems serving as the primary platform for developing and evaluating new methods. On-site

• Available as a full-time or internship opportunity.

Salary range:

$20-$75/hour based on experience. What you will do Responsibilities Develop learning, perception, reasoning, and planning methods for robotic manipulation, including grasping, assembly, tool use, and long-horizon tasks. Explore foundation models, VLM/VLA models, imitation learning, reinforcement learning, and multimodal robot learning. Develop visual and 3D perception methods for object localization, pose estimation, grasp planning, and manipulation. Combine learned robot policies, motion planning, manipulation skills, reasoning, code generation, and tool use into reliable execution systems. Develop feedback, verification, memory, and failure-recovery methods that enable adaptation during physical interaction. Improve policy reliability, generalization, execution efficiency, and inference latency. Design and conduct experiments on physical robots, including automated evaluation, failure analysis, and policy improvement. Contribute to research publications, datasets, demonstrations, and open-source projects. What you bring Requirements MS or PhD in Computer Science, Robotics, AI, Electrical Engineering, or a related field. Strong Python programming skills and solid foundations in machine learning, computer vision, robotics, or related algorithms. Research or engineering experience in one or more of the following areas: robotic manipulation; robot learning; computer vision / 3D vision; multimodal foundation models;

VLM / VLA

models; imitation learning; reinforcement learning; motion planning and robot control; or LLM/VLM reasoning, planning, tool use, or code generation. Experience working with physical robots is highly desirable. Ability to independently implement algorithms, build research prototypes, design experiments, and analyze failure cases. Must be able to start within three months. Must have valid U.S. work authorization; OPT and CPT are accepted. Nice to have Bonus qualifications Experience with dexterous manipulation, contact-rich tasks, bimanual coordination, or force/tactile sensing. Experience with large-scale robot data collection, behavior cloning, diffusion policies, VLA models, or reinforcement learning. Experience with automated policy improvement, memory, feedback, or failure recovery. Relevant publications at major AI, robotics, or computer vision conferences. Strong open-source contributions or demonstrated ability to reproduce and extend recent research.