Biosketch
Sameer Ambekar is a Ph.D. student (Doctoral Researcher) at TU Munich and Helmholtz Munich, fully funded by the Konrad-Zuse School of Excellence in Reliable AI (relAI) and affiliated with MCML, pursuing my Ph.D. under the supervision of Prof. Dr. Julia Schnabel.
Prior to my Ph.D., he completed a Master’s in Artificial Intelligence at the University of Amsterdam and worked as a Research Assistant at IIT Delhi.
relAI Research
Learning to Generalize across Complex distribution shifts in Medical imaging
My research centers on adapting Vision-Language Models (VLMs) through GRPO and test-time reasoning, making models more reliable at inference time. Following a research stay at Stanford University, I focus on test-time thinking for medical imaging, leveraging reasoning models to improve diagnostic accuracy under distribution shift. My work broadly sits at the intersection of RL-based fine-tuning, test-time scaling, and multimodal reasoning.