Projects

  • Programming Languages: C/C++, Java, Python, Assembly, JavaScript

  • Environment: Spring Boot, React, AngularJS, Node.js, Grails

  • Databases: GraphflowDB, Neo4j, PostgreSQL, Oracle, MySQL, Cassandra, MongoDB, Elasticsearch

KTabulator, Research project, University of Waterloo, 2020

  • An interaction system to gather data in a tabular format from Wikipedia and DBpedia.

  • Technologies: React

Linked Open Data for Recommendation System: A Literature Review, CS 848 Project, University of Waterloo, 2020

R2GSync, MMath Thesis project, University of Waterloo, 2020

  • An automatic system to synchronize a GDBMS with the source RDBMS

  • Technologies: Java, GraphflowDB

GraphWrangler, MMath Thesis project, University of Waterloo, 2019

    • A system that allows users to connect to an RDBMS and within a few clicks extract graphs out of their tabular data, visualize and explore these graphs, and automatically generate scripts for their ETL pipelines. GraphWrangler adopts the predictive interaction framework and internally uses a data transformation language that is a limited subset of SQL. A demonstration video can be found here.

    • Technologies: Node.js, AngularJS, MySQL

    • Website

Machine Learning Solution on Join Cost Estimation for Different Join Strategies, CS848: ML for DB, DB for ML, Spring-2019

    • This project aims to predict join cost for different join strategies using neural network with multiple hidden layers in predicting.

    • Technologies: PostgreSQL, Python, Keras, scikit-learn

    • Paper

Detecting Fake News: Towards Making the Internet a Reliable Space , CS680 - Machine Learning for Social Good Project, Winter-2019

    • In this project, I analyze the linguistic features and focuses on full text-based writing patterns of deceptive news. I implement Capsule Network (CapsNet) for fake news detection and show that this model can successfully detect fake news with a maximum accuracy of 87%. I further implement CNN and LSTM models and comparatively analyze the performances with the CapsNet model. I use two different datasets—one with long news articles and another with only short news statements—and investigate the effects of these datasets on our predicting models. I also make a combined dataset to develop a more generalized fake news detection model.

    • Technologies: Python

    • Paper, Poster

Tap or Gesture? Exploring Text Entry Techniques for Smartphones , CS889 - HCI Project, Fall-2018

    • In this research, we evaluated the two most popular and widely used techniques of text input: tap typing and gesture typing. In addition to these two, we also evaluated the hybrid (a combination of tap and gesture typing) technique. We did a controlled experiment to find out which one requires less time when people are in a static position (e.g. sitting or standing) as well as they are on the move (e.g. walking). Apart from that, the experiment also reveals which method people preferred the most for text input.

    • Video Paper Presentation

iPay-Admin, iPay Systems Limited, 2017-2018

    • A backend application to manage the administrative work for the iPay application. iPay is a payment platform in Bangladesh. Through iPay-Admin, the admin users can manage the ipay member, perform member verification, check for fraudulent activities, manage role-permissions of ipay members as well as all iPay admin members, maintain iPay finance etc.

    • Technologies: Java, Spring Boot Framework, MySQL

CRM, iPay Systems Limited, 2017

    • A web application that provides a communication platform between customers and iPay Group. This application gathers all the users complains/feedback/requests in one place so that customer care members as well as iPay admins can contact members through these and take necessary initiatives. iPay admins can also manage their members and control role-permissions through this application.

    • Technologies: Grails, Groovy, MySQL

iPay-Monitor, iPay Systems Limited, 2016-2017

    • A data system aim to capture all the user activities and transactions and perform OLAP queries. Data is collected from different modules through Kafka and stored in MongoDB. We built a backend Node.js application to perform different data analytics, OLAP queries, and automatic monitor alerts from different modules. A web application was dedicated to generate time series graphs, instant updates, and notification on server health.

    • Technologies: Node.js, AngularJS, Angular Material, MongoDB, Kafka

iPay-Web (Front-end web application), iPay Systems Limited, 2016

    • A web application that provides an online payment platform. Through this application, users can make a link to their bank accounts, add or withdraw money to or from the added bank accounts, send money to other iPay members, make payment to merchants, pay bills, etc.

    • Technology: AngularJS

BD ATM Escort, Undergraduate Software Engineering Project, 2014

    • This is an android application which shows the location of ATM booths in Bangladesh with some additional features.

    • Technologies: Android, SQLite

Financiamento, A Personal Finance Management System, Undergraduate Database Project, 2014

    • An online personal account manager through which people can manage their savings, expenditure, keep track of the amount and pay date of the loans, and government taxes.

    • Technologies: Java, Oracle

    • Top ten database project award

E-learning for Autistic Children, Undergraduate Java Course Project, 2012

    • An interactive system for teaching autistic children with attractive graphical user interfaces.

    • Technology: Java

Balloon Shooting Game, Undergraduate C Project, 2011

    • A simple gaming application using IGraphics library of C.

    • Technology: C Programming Language