Avatar

Sumit Vaise

Machine Learning Engineer

[2021 - Present]

Computer Vision Engineer
[2014-2018]

Hello !

Currently working as a Machine Learning Engineer@ Quantiphi Inc. .

I had significant experience in Computer Vision application development through my Computer Vision Engineer (1.5+ yrs in Canada and 4.5+ years in India) and Research Assistant positions.

My whole professional and academic career revolves around pixels, in understanding them, manipulating them and using them for specific purposes through Image Processing and Computer Vision. I developed prototypes, using Python or MATLAB, to understand the underlying concepts and then converted them to an optimized applications using C++.I created application for desktop as well as for the development boards. I also utilizing GCS/AWS for serving model predicitons. I have good experience in Bash scripting and Linux OS. Currently focussing on learning DevOps related technologies.

Interests

  • Image Processing
  • Deep Learning
  • Computer Vision
  • Natural Language Processing
  • DevOps
  • GPGPU Programming

Education

  • MEng in Electrical and Computer Engineering (Course Based), 2019-2020

    Concordia University, Montreal

  • Post Graduate Diploma in Elctronics Product Designing 2013-2014

    Centre of Development and Advanced Computing, Hyderabad

  • BTech in Electronics and Communication Engineering, 2008-2012

    Uttar Pradesh Technical University, Ghaziabad

Skills

Python

Proficient

++

C++

Intermediate

PyTorch

2 years

Deep Learning

3 years

Computer Vision

4 years

Natural Language Processing

< 1 year

Qt

2 years

Docker

< 1 year

Amazon Web Services Instances EC2

< 1 year

Linux

4 years

Django

1 year

Experience

 
 
 
 
 

Machine Learning Engineer

Quantiphi Inc. Canada

July 2021 - Present Montreal, Canada

Responsibilities :

  • The objective of the company isto provide GCP based ML solutions.
    • Technologies used: Machine Learning, PyTorch, Python, GCP, Kubernetes, BigQuery, CI/CD.
 
 
 
 
 

Research Assistant

Concordia University

Jul 2019 – Dec 2020 Montreal, Canada

Responsibilities :

  • Created Project plan and project requirements.
  • Developing Language Models for Non-English languages.
  • Develop speech recognition based deskptop application using PyQt

    • Technologies used: PyTorch, Python and Qt.
 
 
 
 
 

Sr. Computer Vision Engineer

Robert Bosch Engineering and Business Solutions Private Limited

Aug 2017 – Dec 2018 Bangalore, India

Responsibilities :

  • Designed and Developed an end-to-end Deep-Learning based computer vision pipeline to remove densely labelled dataset and convert them into sparsely labelled dataset. The Sparsely Labelled dataset has reduced the training time and computational cost by 32% and is currently used in the project.
  • Developed a 2D cubic shape on car images using OpenCV and Python which further used as a label in training a model.
  • Developed Qt and C++ based semantic segmentation labelling tool.
  • Collaborated with Bosch's Germany team on Autonomous driving project for data visualization and training the VGG network.
  • Project management using tools like JIRA to create, assign, and submit tasks, manage task descriptions and task timelines.

  • Use SVN for version control.

    • Technologies used: PyTorch, Tensorflow, Python, OpenCV, C++, Qt.
 
 
 
 
 

Software Engineer for Image Processing Applicaitons

LnT Technology and Services

Jan 2017 - Aug 2017 Mysore, India

Responsibilities:

  • Developing novel algorithm for stitching two radiographic images using intensity based image registration method.
  • Developed application for human segmentation on thermal images using OpenCV and C++ on RaspberryPi.
  • Developed end-to-end deep learning based multiple object detection application for marine environment. The model was trained ande deployed on NVIDA jetson-TK1 development board. The whole application was automated using SHELL scripting on a Linux operating system.

  • Developed MATLAB based data labelling tool for cloud and mountain segmentation for France ministry of defense. The dataset was further used as an input for Optical flow algorithm.
  • Scrum methodology was followed to keep track of the projects.

Projects

Can you beat me at this?

Rock Paper Scissor Classification.

Multi-Object Detection: Boat, Dock and Human

A multi-object detection for a marine environment running on NVIDIA-JetsonTK1.

Diabetic Retinopathy Analysis using PyTorch

ResNet-101 to classify images, Django for web interface and AWS for deployement of application on cloud.

Person Detection on Thermal Images using OpenCV and C++.

Utilized image segmentation using the thresholding techniques to detect human in the image.

Radiographic image stitching using MATLAB.

Develop a single radiographic image for easy diagnosis.

Certifications

Deep Learning Specialization

Learned about various neural networks like CNN, Sequence Models, Attention Models and Neural Style transfer. This course involved programming of basic architecture and working of CNN, RNN, attention model from scratch.
See certificate

Tensorflow Developer

See certificate

PG Diploma in Electronic Product desigining

Awarded as the batch topper for the course.
See certificate

Contact

  • +15147016472
  • 211, 1975 Maisonneuve Boul W, Montreal, QC, Canada