Portrait of Lekan Molu

Lekan Molu

Senior Researcher

About

Hi there! I am a (Senior) Researcher in the MSR NYC Reinforcement Learning group. I consider myself first an Engineer, and second a Researcher (and certainly not a single-issue researcher) among other attributes. Real-world engineering problems inform my research and in my day-to-day work, I press towards the mark of merging the scientific elegance of machine learning and control theory with the practical impact of robotics as best as I can.

After my PhD and postdoc, and now informed by the safety-critical aspects of complex control, system identification, and automation at scale  — my research thrusts now revolve around safety-critical learning and control in (possibly constrained) complex systems. I now adopt a three-pronged paradigm to problems encompassing game-theoretic, control-theoretic, and learning algorithms with an eye toward the principled characterization of optimality, robustness, and convergence to (collective) equilibrium. I drive my analyses with a full understanding of physical processes backed up by rigorous numerical evaluations and testing on real-world phenomena.

 

Community Service

I am active in the control, AI, and robotics research community, with services cutting  across high-impact  journals/archived publication venues.

Grants reviewer, NatGeo AI for Species Discovery grants.

Associate Editor, IEEE International Conference on Robotics and Automation (ICRA) (2020, 2022).

Reviewer: The International Federation of Automatic Control (IFAC) World Congress,  Automatica, IEEE Conference on Decision and Control (CDC), International Conference on Machine Learning (ICML), American Control Conference (ACC), Conference on Neural Information Processing Systems (Neurips), IEEE/Robotics Society of Japan International Conference on Intelligent Robots and Systems (IROS), Robotics: Science and Systems (RSS), IEEE International Conference on Robotics and Automation (ICRA), The International Conference on Learning Representations (ICLR), IEEE Robotics and Automation Letters (RA-L).