Marcel Hussing

Marcel Hussing is a postdoctoral fellow in the MIT Department of Electrical Engineering and Computer Science, where he develops reliable AI for autonomous control of experimental platforms. Marcel studies reinforcement learning (RL) algorithms that can learn consistently, safely, and effectively in complex environments where each decision is consequential. During his doctoral studies, he worked in the field of replicable RL, producing the first theoretical frameworks and algorithms that enable independently trained RL systems to achieve consistent outcomes while maintaining strong performance. He also advanced methods for stabilizing deep RL, improving robotic manipulation, and enabling autonomous systems to operate under uncertainty. As an MIT-Novo Nordisk Artificial Intelligence Postdoctoral Fellow, Marcel will develop reliable autonomous learning frameworks that combine RL, robotics, and scientific experimentation to accelerate discovery in the physical sciences and engineering. His work has the potential to empower autonomous lab systems and robots that generate new scientific insights while advancing trustworthy AI for high-stakes applications across research, industry, and society.