About Me

Rishav Chourasia

I am a first-year Ph.D. student at School of Computing, National University of Singapore under the supervision of Prof. Bryan Kian Hsiang Low. As part of my Ph.D. research work, I am interested in studying benefits of knowledge sharing among group of learners (bandits, RL agents or basic function approximators) and finding ways to leverage it for speeding up learning. Non exhaustive range of topics I am interested in : Social Choice Theory, Multi-agent Learning, Reinforcement Learning, Optimization algorithms.

My Journey

Ph.D. in Computer Science

As part of MapleCG lab, I am being advised by Prof. Bryan Kian Hsiang Low.

Aug 2019 - Present
School of Computing, National University of Singapore, Singapore

Student Developer in Google Summer of Code

Contributed to an experimental Compiler for Chainer ML framework.

May 2019 - Aug 2019
Chainer, NumFOCUS organization, GSoC 2019

Data Scientist Internship

Worked on demand prediction for assisting Eat.Fit kitchens’ inventory procurement.

Apr 2019 - Jun 2019
Cure.Fit, Bangalore, Karnataka, India

Research Fellowship

As part of Machine Teaching group, I explored application of Social Choice Theory in context of Knowledge Sharing in Q-learning.

Nov 2018 - Feb 2018
Max Planck Institute for Software Systems, Saarbrucken, Germany

Software Development Engineering

Helped maintain and develop an amazon wide fraud detection service as part of Transaction Risk Management Team.

Jul 2018 - Nov 2018
Amazon Development Center, Bangalore, Karnataka, India

Software Development Intern

Developed platform for tracking and modifying configuration changes in fraud check pipelines.

May 2017 - Jul 2017
Amazon Development Center, Bangalore, Karnataka, India

Research Internship

Explored efficient type-reduction techniques for Multidimensional Type-2 Fuzzy sets.

May 2016 - Jul 2016
Hanyang University, ERICA campus, Ansan, South Korea.

Robotics Club Manager

Organized events, conducted hackathons, guided student projects.

Mar 2016 - Jul 2017
Robotics Club, Indian Institute of Technology, Guwahati, Assam, India

B.Tech in Computer Science and Engineering

Bachelor Thesis: Optimal Swarm RL: An Improved Deep Exploration Strategy.

Jul 2014 - Jun 2018
Indian Institute of Technology, Guwahati, Assam, India

My Projects

Extending Elichika: Support for continue, break, return and more

I am helping improve an experimental compiler for transforming Chainer ML models to Open Network Neural Exchange (ONNX) representation. My contributions so far include support for python dictionary datatype, python syntax like continue, break, with, return, etc. and support for python builtins such as len(), sum(), etc.

...   ...

Optuna

I am helping improve Optuna, a hyperparameter optimization framework. My contributions include suppoort for MXNet pruning callbacks and examples for PyTorch, Tensorflow, MXNet and Keras.

...   ...

Super Mario Bros AI

Super Mario AI is an epsilon soft policy based Q-learning model that learns to plays the famous Super Nintendo Entertainment System game Super Mario Bros. I used a Lua scriptable SNES simulator to create the environment for RL models.

...   ...