I am a Research Engineer in AI and Robotics at Shadow Robot, where I work on the co-design of learning algorithms and robot embodiment for dexterous manipulation. I completed my PhD at the University of Edinburgh under the supervision of Prof. Michael Mistry.
My research lies at the intersection of embodied AI, robot learning, and dexterous manipulation. I am motivated by building intelligent robotic agents that can perceive, reason, and act effectively in contact-rich real-world environments. In particular, I aim to enable robots with human-level agility and intelligence, and even beyond-human dexterity, for challenging manipulation tasks.
Research Focus
My work focuses on multimodal robot learning, data-efficient imitation and reinforcement learning, long-horizon dexterous manipulation, and sim-to-real transfer for reliable deployment on real robot.
- Multimodal robot learning for dexterous manipulation, especially with tactile sensing
- Data-efficient imitation learning from limited demonstrations
- Efficient reinforcement learning for long-horizon dexterous tasks with sparse rewards
- Sim-to-real transfer for robust real-world deployment
- Co-design of robot learning algorithms and dexterous robot morphology
- Cross-embodiment learning and morphology evolution
Selected Publications
