Agentic Robotics Research Scientist

Summary of role: This role focuses on developing next-generation agentic robot learning systems that enable robots to reason, plan, learn, and autonomously execute complex, long-horizon tasks in the real world. The researcher will work at the intersection of foundation models, robot learning, planning, world models, and embodied intelligence, translating cutting-edge AI research into robust capabilities on physical robot systems. The role involves both fundamental research and end-to-end system development, with opportunities to publish at top-tier venues and shape the technical direction of agentic robotics.

What you'll be doing:

• Conduct research on agentic robot learning systems that enable robots to reason, plan, learn, and complete long-horizon tasks in real-world environments.
• Develop methods spanning foundation models, vision-language-action models, task and motion planning, reinforcement learning, imitation learning, world models, memory, and tool use.
• Explore how autonomous agents can perceive their environment, decompose goals, make decisions, recover from failure, and improve through interaction.
• Work closely with robotics, computer vision, control, and platform teams to deploy research prototypes on physical robot systems.
• Design and run experiments in simulation and on real hardware; analyse results and iterate rapidly on system performance and reliability.
• Contribute to the team’s technical research direction, including identifying promising research questions and turning them into practical capabilities.
• Publish research outcomes in leading conferences and journals, and contribute to patents or open-source work where appropriate.

What we're looking for

• PhD in Robotics, Computer Science, Artificial Intelligence, Machine Learning, or a related field.
• Strong 1st-author publication record at recognised top-tier conferences or journals, such as NeurIPS, ICML, ICLR, CoRL, RSS, TRO, IJRR, ICRA, IROS, CVPR.
• Strong understanding of modern AI and robotics methods, with experience in one or more areas including embodied AI, reinforcement learning, imitation learning, planning, large language models, vision-language models, world models, or multi-agent systems.
• Strong programming ability in Python and familiarity with modern machine learning frameworks such as PyTorch, JAX, or TensorFlow.
• Demonstrated ability to independently drive research from problem definition through experimentation, evaluation, and communication of results.
• Curious, rigorous, and comfortable working on open-ended research problems in a fast-moving environment.

Preferred qualifications

• Research experience deploying robot learning systems on physical robots.
• Research experience with dexterous manipulation and humanoid robotics.
• Experience integrating LLMs or vision-language models with planning, control, memory, simulation, or real-world execution.
• Familiarity with robotics simulators such as Isaac Sim, MuJoCo, Habitat, ManiSkill, or equivalent platforms.