K-M Samiul Haque

K-M Samiul Haque

Full Stack Developer

Work Experience

Full Stack Developer - University of Toronto
Sep, 2017 - Present

  • Developed a scalable web application using Django, Bootstrap, jQuery, and PostgreSQL for tracking employee hours and enrollment trends for several departments.
  • Created historical models and data visualization using Django and ChartJS, which was used to decide on future course offerings.
  • Created real-time updates for concurrent operation by multiple users using Django-Channels, Redis and WebSockets.
  • Created a custom excel parser for parsing weekly reports into JSON format for use by back-end functions.
  • Built custom back-end functions for calculating quotas, TA hours, finances and other departmental affairs.

QA Automation Engineer - RBC
May, 2017 - August, 2017

  • Developed a Test automation web portal and QA documentation site using the Django web framework, MongoDB, Django-Channels, Redis, jQuery, Jenkins and Nginx for concurrent execution, viewing, and logging of automated scripts with real time data analytics. Saving the department $100,000+ per year, while reducing execution time by 80%.
  • Developed a web based database viewer for MongoDB using Django and Bootstrap.
  • Migrated test data stores from Excel and PostgreeSQL to MongoDB.
  • Responsible for creating automated test cases via Python scripting and the selenium library.
  • Executed and developed automated test scripts using C# and the .NET framework.
  • Built custom scripts in C# to interface with the IBM Personal Communications emulator and perform mainframe testing using the LeanFT libraries.

Back-end Developer (remote) - Kabita Inc.
May, 2016 - August, 2016

  • Developed an e-commerce CRUD application using Bootstrap, Django, and MongoDB.
  • Migrated the entire back-end to cloud-based solution allowing rapid scaling.
  • Containerized application using Docker to ensure rapid deployment.
  • Developed applications under an Agile environment to ensure concurrent testing of all features.

Latest Projects


RAWCast

Created a Desktop application using C# and the .NET framework to cast Canon .CR2 RAW image files to nearby Google Chromecasts, converting them on the fly using a custom C image manipulation library.

The user first selects a folder which contains the Canon RAW image files, the application then decodes every image in the folder, generating a slideshow preview of the selected images. When an image is chosen to be casted the application converts the image to a commonly used image format and serves it to nearby Chromecasts. This is especially useful for photographers who want to cast their RAW image files to their TV.


MirrML

MirrML is a web application built on the Flask web framework that analyzes user's clothing given a set of pictures of the users clothing and return what type of attire it is.

It uses the Clarify Image Recognition API and a custom model we trained using a web scraper to analyze any picture of the users attire and return a probabilistic match, regarding what kind of style it is. This may be business casual, formal, or evening. It also matches the user's clothing with their friends and check which of their friends have the most similar clothes.


LendR

NFC based social micro-financing application

LendR is an android application that uses NFC technology to loan money to friends who you trust. Allowing fast, seamless and secure money transfers at a moment’s notice.

We even built a score system to differentiate users who pay back their loans on time, from defaulters, we call this system the "karma system". If a user chooses to not pay back the lender on time, they accumulate interest on their loan as well as negative karma points. With each decreasing karma point, the users borrowable amount decreases as well until it reaches a threshold and cannot borrow anymore.

A custom backend was also setup to keep track of users’ karma points and transactions, as well as a custom prompt was added to ensure lenders can verify the borrowers’ karma rating before authorizing the transfer.


Ultralux

Laser based early warning system

A IoT based application that serves as an early warning system, sending tweets using the twitter API via lasers and solid state switching by a Raspberry Pi to a android receiver.

The data is compressed, converted into binary, and then is transmitted via modulated laser flashes toward the android camera. The android app receiver then receives this data and converts the binary data back into Unicode.

My Git Repository

Loading Chart....

Favourite coding music