Agentic Physical AI | AI Agents for Robotics

My work is focused on the design and implementation of agentic AI systems in robotics: agents that are more autonomous, that can plan, perceive, and act in physical spaces, under uncertainty, with partial information, and adapt to changing conditions. I bring expertise in:

• Vision-Language-Action models, foundation models for robotics: integrating perception, language understanding, high level planning and low-level control.

• Task and goal hierarchies, long-horizon planning, subgoal decomposition, embodied chain of thought.

• Handling safety, reliability, real-world constraints (energy, latency, physical safety, human interaction).

• Transfer from simulation to real robots; generalisation across embodiments; modular architectures for agents.

I have co-authored recent publications in embodied agent planning; explored models that generalise across environments and tasks; experience with both software stack and real robotic testbeds. Looking to engage in projects developing generalist robotic agents, interactive robot assistants, and adaptive physical agents in unstructured environments.