About Myself

Interested in solving problems related to machine learning, data and cloud computing. I spend most of my time updating my skills and keeping my self aware of the latest trends in machine learning and software systems. I also spend a significant amount of time doing solo explorations, on foot and over the internet.

Ongoing Projects:

Energy Usage Prediction: Predicting top Energy demand days by implementing Time Series models like ARIMA and complex models like RNN on IESO Hourly Ontario Demand Time Series Data.
Interpretable Deep Learning: Image Pre-processing using Bit-Plane Manipulation to Improve the Robustness of CNN based Image Classification Systems against Adversarial Examples.
GPU Programming: Speeding up Back Propogation algorithm using CUDA.

2+ years

Industry Exp.

5+ years

Academic Exp.

2

Research Papers

Python 80%

C/C++ 85%

Java 85%

MySQL 80%

Docker & Kubernetes 70%

Tensorflow 80%

Linux 85%

GitHub 80%

OpenStack 80%

Projects and Experiments

CUDA Programming

Code to optimally utilize the resources on GeForce GTX 980 and obtain significant speedup concepts like pinned memory and padding to name a few.

Image Enhancement

Implementing concepts like log and gamma transformation including image equalization in matlab.

Portfolio Optimization

Implemented Markowitz Mean-Variance Portfolio Theory in Matlab for the stocks returns of key investors in Blockchain Technology.

Deep Learning (CNN)

Performed analysis on regression models and deep neural networks for various regularization and initializations techniques in Tensorflow.

Machine Learning Pipeline

Deployed an environment to periodically scan images from Google Bucket and generate inference on a large scale using Kubernetes & GCP.

Java WebService

Deployed an end-to-end framework and trained a CNN using Transfer learning, to classify fruits and vegetables based on Image and Weight.

Trading Algorithm

A moving-average based algorithm tested on Rotman Interactive Trader (RIT) application.

OpenStack

Implemented a monitoring solution for OpenStack deployment using Nagios. This resulted in a research paper and further research using Sensu.

AMP Cisco

Was responsible for testing and validation of Advanced Malware Protection Framework on Redhat machines while working at Cisco Systems.

UCS Cisco

While working at Cisco Systems, I got an opportunity to setup a virtual environment from bottom up using Cisco Unified Computing System.

Data Load Utility

A library to generate load by analyzing MySQL database schemas. Usage of this library would be to synthesise arbirtary values for stress testing.

Autonomous Underwater Vehicle

A literature and experimental survey of software systems available for robots.

Natural Language Processing

Analysis of which English verbs can be borrowed into Indian languages and what factors influence the borrowing and frequency of usage.

Smart Library

A working prototype of Smart Shelf using RFID system, capable of generating the location, orientation, and pick count of books in the shelf.

  • Oct 2017 - July 2018

    HyperWorks Imaging

    Software Researcher
    Bangalore, India

  • Aug 2016 - Sept 2017

    Cisco Systems

    IT Engineer
    Bangalore, India

  • June 2015 - July 2015

    ThoughtWorks

    Software Engineer
    Bangalore, India

  • Sept 2018 - Present

    University of Toronto

    Graduate Student
    Computer Engineering and Analytics

  • Aug 2012 - Aug 2016

    RV College of Engineering

    Undergraduate Student
    Computer Science and Engineering