(the second LinkedIn directs
to our non-profit)
Srikanth Chandar
AI Research
Animal Welfare
Hi there, welcome to my website! I try to include all my projects, research, and animal welfare work here.
I am currently a PhD student at Johns Hopkins Center for Alternatives to Animal Testing (CAAT), focusing on computational biology. Previously, I was an ML Ops Engineer at AT&T Labs, having graduated from Columbia University with an MS in Electrical Engineering. Prior to this, I was an AI Engineer at Intel Corporation, having graduated from PES University. My latest research interests are in Foundation Models and Bio/Chem informatics.
A little about the non-technical side of me- I feel very strongly about animal abuse. I help run Ground Animal Welfare, which promotes animal welfare initiatives through campaigns, volunteering activities, and even technical projects.
Research Interests
Anything ML excites me, and I like taking on a challenge that could shape a new paradigm in the same space, be it application-based or research. My preference in Applied Research is in areas like Bio/chem informatics, Deep Learning, and Data Science. I am exploring my finer interests in these subjects, one project at a time, each on a different topic from the previous.
My recent research inclinations have been in Foundation Models and how they may forecast time series data. My work at AT&T Labs involves the same. Federated Learning (FL) has been another focal point of my research lately, and my work at Intel included making FL more communication-efficient.
Publications
Chandar, Srikanth, et al. "Communication Optimization in Large Scale Federated Learning using Autoencoder Compressed Weight Updates." arXiv preprint arXiv:2108.05670 (2021).
Chandran, P., Bhat, R., Chakravarthi, A., & Chandar, S. (2021). Weight Divergence Driven Divide-and-Conquer Approach for Optimal Federated Learning from non-IID Data. arXiv preprint arXiv:2106.14503.
Chandar, Srikanth, et al. "Road Accident Proneness Indicator Based On Time, Weather And Location Specificity Using Graph Neural Networks." arXiv preprint arXiv:2010.12953 (2020).
APA