Jeevan Devaranjan

I'm a masters student at the University of Toronto working with Sanja Fidler. My general research interests span Computer Vision, Reinforcement Learning, Generative Modelling and Control along with their use in simulating large systems. I did my undergrad at the University of Waterloo majoring in Computer Science and Combinatorics & Optimization (CO). Aside from research I've also worked as a software engineer with experience in distributed machine learning and scaling up existing ML deployments. I've always been interested in all things mathematics ranging from functional analysis to combinatorial enumeration. In my highschool days I used to be active in math/programming contests and general problem solving (used to be somewhat active on math stack exchange back then).

Email  /  CV/Resume  /  Google Scholar  /  GitHub  /  LinkedIn

Education

MSc. Computer Science

BMath. CS and C&O

Research

Meta-Sim2: Unsupervised Learning of Scene Structure for Synthetic Data Generation

Work Experience

Quantitative Researcher

Using math to make money.

Research Scientist Intern

Working on traffic modelling with Multi-Agent Reinforcement Learning with Sanja Fidler and Jonah Philion. One paper under review for CVPR 2023, another to submitted for ICCV 2023.

Machine Learning Research Intern

Research on unsupervised 3D pose estimation with Satya Gorti and Maksims Volkovs

Deep Learning Research Intern

Generative Reinforcement Learning and Semi Supervised Synthetic Content Generation research with Professor Sanja Fidler and Amlan Kar.

Software Engineering Intern

Added support for probabilistic models on the ML workflow pipeline using GraphQL. Implemented probabilistic ML models such as Deep Markov Models and Bayesian CNNs in the Tensorflow and Torch Libraries using Edward, TF Probability and Pyro.

Security Developer

Worked on developing software for logging and handling security events and exceptions.

Design and source code from Jon Barron's website via a jekyll fork from Leonid Keselman