Home
Hello 
,
it’s great that you found me here! My name is Marcel Hussing and I am a postdoctoral researcher in the Improbable AI Lab at MIT, where I work with Pulkit Agrawal. I am also an MIT–Novo Nordisk AI Postdoctoral Fellow. My postdoctoral research explores how learning algorithms for sequential decision making can help solve scientific problems.
I completed my PhD in computer science at the University of Pennsylvania, where I was advised by Eric Eaton as a member of the Lifelong Machine Learning group. I am broadly interested in what I call reliable sequential decision making, and my research has particularly focused on stability in reinforcement learning.
My work ranges from learning theory and neural network training to deploying machine learning on real hardware. At Penn, I had the great fortune to collaborate with and learn from Michael Kearns, Aaron Roth, CJ Taylor, Pratik Chaudhari, and many others.
During my PhD, I was a research intern at Microsoft Research New York, where I worked with John Langford on abstract planning for goal-conditioned transformers. I also interned at Meta FAIR with the JEPA team, including Mido Assran and Scott Fujimoto, working to bridge the gap between world models and robotic control.