Summary
Debajyoti Sengupta is a Postdoctoral Scientist at EPFL’s MIP:Lab with nine years of experience at the intersection of machine learning, physics, and astrophysics. He develops computational frameworks for generative AI and video analysis, linking visual dynamics to neural activity and pushing practical advances in score-based diffusion and transformer-based generative models. During his PhD at the University of Geneva he reduced generation compute by 10x while improving fidelity and built weakly supervised tagging methods that boosted astrophysical signal detection and dataset purity. His work spans GANs for calorimeter simulation to anomaly detection in stellar stream hunts, blending theoretical rigor with production-minded efficiency gains. Based in Geneva, he combines deep research credentials with hands-on model engineering and a knack for turning noisy, real-world scientific data into actionable insights. Off-hours he balances intense problem-solving with a reflective streak—“sometimes I sits and thinks, and sometimes I just sits”—suggesting a thoughtful, observant approach to complex challenges.
9 years of coding experience
Bachelor of Science (B.Sc.), Bachelor of Science (B.Sc.) at Scottish Church College
Junior school, Junior school at D.A.V public school, Gua
Master of Science - MS, Master of Science - MS at University of Calcutta
B.D.M. International
Doctor of Philosophy - PhD, Doctor of Philosophy - PhD at University of Geneva