Exploring how learning systems can adapt, coordinate, and move from simulation into the real world.
I work at the intersection of reinforcement learning, multi-agent systems, robotics, and control.
My most recent published work is ADiCo, a framework that combines multi-agent reinforcement learning with Extremum Seeking Control to automatically adapt behavioral diversity based on task performance.
My current work explores how these ideas can be applied to robotic systems, sim-to-real learning, and autonomous multi-agent decision-making.
Multi-Agent Reinforcement Learning · Robotics · Adaptive Control
Robot Learning · Sim-to-Real · Autonomous Systems
Strategic AI · Emergent Multi-Agent Behavior
🔬 ADiCoAdaptive behavioral diversity in cooperative multi-agent reinforcement learning using Extremum Seeking Control. |
Taking learned multi-agent policies from simulation toward ROS2 and robotic systems. |
|
Learning robotic manipulation trajectories from demonstrations. |
Exploring strategic decision-making and skill transfer in StarCraft II. |

