Karan Shah

I am a Computer Scientist at the Deutsches Zentrum für Luft- und Raumfahrt (German Aerospace Centre), where I build reinforcement learning and LLM-driven agents for autonomous navigation of surface vessels in complex marine environments. My work combines learning-based perception with environment-aware control policies to enable real-time decision-making in simulation. Before DLR, I did my M.Sc. in Informatics at the Technical University of Munich.

My research interests include crowd modelling and simulation, dynamical systems, chaos and number theory. I am fascinated by the distribution of prime numbers and the Goldbach’s conjecture.

I have worked with AUDI AG, BMW GROUP AG, CodeChef, Quantiphi Inc. and NSE (National Stock Exchange of India Limited) in various capacities on projects in the domains of Data and Business Analytics, Machine Learning, Autonomous Driving and Game Development. I’ve also worked with the Chair of Traffic Control and Engineering at TU Munich on the SAVeNoW project where I conceptualized and designed a digital twin of the city of Ingolstadt in Germany.

I volunteer for 15X4 Munich, where we organize free educational meetings, competitions, industrial visits and other events. My goal is to promote the idea of science communication and make knowledge more approachable.

You will usually find me playing chess on my phone in a corner at a party. Up for a game of chess? Challenge me here.