I’m Saptarshi, a Ph.D. candidate in Statistics at UC Berkeley, advised by Prof. Peter Bartlett. Before joining here, I completed both my Bachelor of Statistics and Master of Statistics at the Indian Statistical Institute, Kolkata.

Equipped with a strong mathematical foundation I’m interested to explore the depths of statistical learning, especially deep learning and its real-world applications. My current research is primarily focused on exploring and understanding why deep learners work from a statistical viewpoint. My doctoral research explores the behavior of different deep learners, especially, generative models under the so-called “manifold hypothesis”. If you’d like to know more, check out my current works on GAN, WAE, federated and general supervised learners.

In my free time I enjoy hiking, reading novels and cooking exotic dishes.

Education

(2020 - Present) University of California, Berkeley
             Ph.D. in Statistics
             Advisor: Prof. Peter Bartlett
(2018-2020) Indian Statistical Institute, Kolkata, India.
             Master of Statistics (M.Stat), first division with distinction
             Specialization: Computational Statistics and Applied Statistics.
             Advisor: Prof. Swagatam Das
(2015-2018) Indian Statistical Institute, Kolkata, India.
             Bachelor of Statistics, Honours (B.Stat, Hons.), first division with distinction