Home About Publications Projects Resume Contact
Sourav Biswas
Student, Leader, Creator

About Me

I'm in my final year of undergraduate studies at the University of Waterloo pursuing Computer Science and Finance. I have interests in machine learning and data privacy.

I'm currently a research assistant under the supervision of Professor Gautam Kamath.


Upcoming research internship at Nvidia AI lab supervised by Professor Sanja Fidler.



Previously I researched LiDAR compression methods with deep learning for autonomous vehicles at Uber Advanced Technologies Group supervised by Professor Raquel Urtasun.



I also have experience as a software engineering intern at Wish on the product web development team.



I was also a quantitative analyst at BMO Capital Markets where I created financial tools to analyze bond markets.

Aside from mathematics and computer science, I invest my time in basketball, track & field, weightlifting, chess, drawing, reading and poker.

Research Publications

MuSCLE: Multi Sweep Compression of LiDAR using Deep Entropy Models
Sourav Biswas, Jerry Liu, Kelvin Wong, Shenlong Wang, Raquel Urtasun Advances in Neural Information Processing Systems 33 (NeurIPS 2020).
[paper link]

CoinPress: Practical Private Mean and Covariance Estimation
Sourav Biswas, Yihe Dong, Gautam Kamath, Jonathan Ullman
Advances in Neural Information Processing Systems 33 (NeurIPS 2020).
[paper link] [code]


Hover for details or click on a project for more information.
More projects are listed on GitHub
Computer Vision Weightlifting Coach
Built ML models to analyze weightlifting videos for correct posture. Feeds data from MPII pose estimation with OpenCV.
Built with: Python, Jupyter Notebook, Scikit-Learn, OpenCV, OpenPose
Sit back and relax in your next meeting or lecture. This web application will take the notes for you!
Built with: Python, HTML/CSS, JavaScript(jQuery), Flask, SQL
A full stack Ethereum decentralized application to serve as a fundraising platform for charities and crowdsourcing.
Built with: Truffle (Ganache), MetaMask, HTML/CSS, JavaScript(jQuery), Solidity
NBA Predictive Analytics
Analyzes NBA players' shooting strengths and habits. Uses machine learning to simulate an athlete's gameplay.
Built with: Python, Scikit-Learn, Pandas, Matplotlib


sourav.j.biswas [at] gmail [dot] com



Google Scholar