Sourav Mishra

Senior Machine Learning Engineer

theAstate-Kinliser

I work on computer vision, graphics and machine learning at theAstate, a subsidiary venture of Kinliser Japan. I manage projects which range from image-aided estimations to generating immersive 3D layout. I have recently forayed into LLMs & designing bespoke solutions using it.

I earned my PhD from University of Tokyo in 2021 in limited-data deep learning. I am also an alumnus of Virginia Tech where I earned dual Masters in Electrical & Computer Engineering (2013) and Biomedical Imaging (2015).

Profile

ML Interests

3D and Computer Vision

Automating visual assessments, fine-grained image classification, limited data learning

AI in Healthcare

Applied machine learning in healthcare

NLP & LLMs

Foundation models, RAG and vector search

Curriculum Vitae

Education

University of Tokyo

Ph.D, Computer Science

University of Tokyo, 2021

Virginia Tech

M.S., Biomedical Imaging

Virginia Tech, 2015

Virginia Tech

M.S., Electrical & Computer Engineering

Virginia Tech, 2013

Manipal University

B.S., Electronics & Biomedical Engineering

Manipal University, 2009

Experience

Senior ML Engineer

theAstate-Kinliser, 2022-Present

Senior Data Scientist

Anymind Japan, 2021

Instructor

Amity University, 2015-2016

Selected Publications

Robustness of Deep Learning Models in Dermatological Evaluation: A Critical Assessment

S Mishra, S Chaudhary, H Imaizumi, & T Yamasaki

IEICE Transactions on Information & Systems

Vol 104 (3), pp. 419-429, 2021

Interpreting fine-grained dermatological classification by deep learning

S Mishra, H Imaizumi, & T Yamasaki

CVPR-W 2019 (Selected for Oral)

Improving image classifiers for small datasets by learning rate adaptations

S Mishra, H Imaizumi, & T Yamasaki

MVA 2019 (Selected for Oral)

SLATE: Virtualizing multiscale CT

S Mishra, KS Sharma, SJ Lee, EJ Fox & G Wang

Journal of X-ray Science & Technology

Vol 20 (2), pp. 239-248, 2012

Collimator width optimization in X-ray luminescent computed tomography (XLCT) with selective excitation scheme

S Mishra & R Kappiyoor

Journal of Medical Imaging & Health Informatics,

Vol 4 (5), pp. 681-686, 2014

Contact

Email

sourav [at] hal.t.u-tokyo.ac.jp

Computer Vision & Media Lab

Engineering Building 2

University of Tokyo

7-3-1 Hongo, Bunkyo-ku

Tokyo 113-8656

Email

sourav [at] astate [dot] ai

theAstate-Kinliser

12F, Tokyo Toranomon Global Sq.

1-3-1 Toranomon, Minato-ku

Tokyo 105-0001