![]() |
Senior ML Engineer 3D & Computer vision Contact sourav [at] astate [dot] ai Pronoun: He/Him Short Bio CV |
Permanent Snailmail 〒113-8656 Computer Vision and Media Lab Engineering Building 2 The University of Tokyo 7-3-1 Hongo, Tokyo |
My interests lie in computer vision, graphics and machine learning. I am currently working at theAstate, a startup backed by University of Tokyo's incubator. I lead the efforts on applying 3D computer vision techniques for immersive indoor layouts. I have worked previously for Anymind Japan, implementing computer-vision centric ad-tech pipelines & recommender systems. I completed my PhD in EECS (2021) at the University of Tokyo. My doctoral research was about optimization and deep learning in very homogeneous, small-data regime. I have dual M.S [Electrical & Computer engineering (2013) and Biomedical imaging (2015)] from Virginia Tech.
My PhD research focused on effective learning from noisy, small and typically imbalanced datasets. My techniques found significant use cases in healthcare. Our research group produced several state-of-the-art results for anomaly detection in medical image datasets (dermoscopic and user-submitted photographs for telemedicine in dermatology). In addition to ML, I have dabbled in computational imaging and photogrammetry. I have been generously supported in the past by the following: Microsoft Research (2018), MEXT fellowship (2017-2020) at the University of Tokyo, Bradley scholarship (2012) at Virginia Tech & Stanford Biodesign fellowship (2010).