Bio
I am a Research Associate at Google Research working in the Mixed-Mode User Understanding team, under the mentorship of Dr. Gaurav Aggarwal. I am currently working to build fully input-adaptive and dynamic deep neural networks for fine-grained visual recognition.
Before this I was a Data Scientist at Flipkart in the Trust & Safety team supervised by Dr. Shourya Roy, and previously by Dr. Jatin Thukral. Prior to joining Flipkart, I completed my M.S (by Research) at IIT Madras advised by Dr. Mitesh M. Khapra and associated with RBC-DSAI lab (led by Dr. Balaraman Ravindran). I have completed my undergrad/B.Tech. in Computer Science from Guru Nanak Dev University.
My research interests broadly lie in Computer Vision, Explainable Aritifical Intelligence and Model Compression. I strongly believe that fundamental research in these fields for sensitive use-cases (e.g. AI for criminal justice, automated health diagnosis etc.) can go a long way in enhancing human trust-worthiness of AI models.
My CV is available here.
Preprints and Publications
Efficient Video Classification Using Fewer Frames
CVPR 2019 Conference (oral presentation)
Paper
Poster
Studying the Plasticity of Deep Convolutional Neural Networks Using Random Pruning.
Journal of Machine Vision and Applications 2019
Paper
Poster
Masters Thesis (IITM): Efficient Video Classification
Advisor: Dr. Mitesh M. Khapra
Thesis
Slides
|