Avatar

Sunil Kumar

Graduate Research Assistant

University of Victoria

About Me

I am a Master of Computer Science Student at the University of Victoria. I have 3 years of software development experience with software product companies including Drishti-Soft Solutions aka Ameyo and Zillious Solutions in India. I have worked on new stand-alone product development projects as well as projects in a team for high priority customer, this sometimes includes understanding customer requirements directly from customers. As I am a passionate gamer, I would love to be part of game development industry.

Interests

  • Machine Learning
  • Reinforcement learning
  • Game Development

Education

  • M.Sc. in Computer Science, 2019 - present

    University of Victoria, Canada

  • B.Tech in Information Technology, 2014

    National Institute of Technology, Durgapur, India

Skills

Python

PyTorch, Numpy

Java

Hibernate, JDBC, GWT, Spring Boot

JavaScript

SQL

MySQL, PostgreSQL

Machine Learning

Deep Learning, Reinforcement Learning

Web Development

HTML, CSS, PHP, JSP, AngularJS, Bootstrap

Other Programing Languages

Golang, C, C++, Groovy

Agile Practices

Scrum, Distributed Scrum

Tools & Frameworks

Docker, Kubernetes, gRPC, REST, Git, Electron, Jasper, EMF

Experience

 
 
 
 
 

Technical Associate

Zillious Solutions

Jul 2018 – Dec 2018 *New Delhi, India*
  • Developed a solution for a long-standing issue with java enum reaching its size limit, with minimum changes in the application.
  • Added support for customizable vouchers and integrated multiple payment gateways including Razorpay, Paytm, etc.
  • Integrated a pre-developed machine learning model into the application for finding duplicate hotels. Decreased the time for marking duplicate hotels by 90% in live search results for hotels.
 
 
 
 
 

eSports Player

Ragequit eSports Solutions

Jun 2016 – Jan 2018 India
  • From Dec 2016 to June 2017 was part of DOTA 2 ESports team Sponsored by Ragequit eSports Solutions, New Delhi.
  • Runner up in COBX National DOTA 2 tournament (Buriza Team, Hyderabad, May 2017).
  • Won Playtonia DOTA 2 tournament (Nefarious Team, Delhi, Aug 2017).
  • Won Indian LAN Gaming Cup DOTA 2 tournament (Nefarious Team, Delhi, Oct 2017).
  • Runner up in Aorus Land DOTA 2 National tournament (Nefarious Team, Bangalore, Jan 2018).
 
 
 
 
 

Software Developer

Drishti-Soft Solutions

Jun 2014 – Sep 2016 Gurugram, India
  • Awarded with Excellence Award 2016 and won /tmp 2014 a 10 days 24x7 coding event.
  • Majorly worked on Ameyo Report, a GWT based front-end, and Jasper based backend; This includes a UI renew, backend enhancements and automating the generation of Transformers, a data de-normalization component, by using EMF modeling.
  • Created a report scheduler based on the Quartz library with support for centralized reporting and graphical reports.
  • Increased performance for a large customer’s reporting database by 90% using table partition and multicolumn indexes.
  • Increased performance of a statics managing component by 60% using soft references based cache for reflective calls.
  • Developed a context-aware report designer with the Front-end based on GWT with data populating based on EMF model instances and outputs a proprietary file format (.drxml) which converted by the backend into a JasperReports based object. It was used to give demos to international customers as a future project.
  • Created Magic Numbers: Predicts hardware requirements (CPU, RAM, and HDD) based on customer functional requirements.
  • Delivered components for the automation of the collection process for an Indian Bank. Implemented a fresh UI, Integrating Hitachi Content Platform (HCP) for report storage, created a custom SSO authentication server & various domain-specific features like Centralized Exclusion.
  • Interviewed candidates, prepared questions and did the training for new employees for the 2016 and 2017 hiring seasons.
 
 
 
 
 

Software Developer Intern

Morphing Machines

Jun 2013 – Aug 2013 Bangalore, India
  • Designed and developed the front-end interface for a research project called ‘Development of A Flexible Framework for A Network On Chip’. We wrote a research paper about it (git.io/Jen9l).

Projects

Dota 2 Draft Based Win Prediction

train a machine learning model to predict the winning team for the popular online game Dota 2

Demo Matchmaking using open-match

Final 10 player deathmatch matchmaking demo’s docker images were deployed on a single Kubernetes cluster on Google Cloud Platform and analized using Grafana and Prometheus

DQN Agent playing Coinrun

A pytorch based implimentation of DQN and DDQN to play coinrun game

Contact

  • University of Victoria, Victoria, BC, V8P 5C2, Canada
  • Skype Me