Skill Set


  • Programming Language C/C++, Python, Java, Javascript, SQL

  • Machine Learning Modules scikit-learn, numpy, xgboost

  • Deep Learning Frameworks PyTorch, Tensorflow, Keras, Caffe

  • Data Analysis and Visualization Pandas, matlab, R, matplotlib, opencv

  • Web Frameworks Ruby on Rails, PHP

  • Database MySQL, PostgreSQL

  • Project Management Tools Git, Slack, Basecamp


Employment History

Advanced Mobility Analytics Group, Senior Software Engineer - Computer Vision and Video Analytics (Part-Time) [Feb 2021 – Current]

  • Lead the development and implementation of new video analytics (VA) & artificial intelligence (AI) features within the platform

  • Adopt and apply cutting-edge technologies to improve VA & AI performance

  • Work with and mentor junior VA & AI engineers within AMAG


Queensland University of Technology, Sessional academic [Mar 2019 – Jun 2020]

Teaching the following courses to QUT undergraduate and master’s students.

  • CAB 320 – Artificial Intelligence, Semester 1, 2020

  • IFN 680 – Advanced Topics in Artificial Intelligence, Semester 2, 2019

  • CAB 320 – Artificial Intelligence Semester 1, 2019

  • I teach students how to use deep learning and machine learning frameworks (TensorFlow, Keras, scikit-learn) and algorithms using Python programming language.


Tiger IT Bangladesh Ltd, Software Engineer

[Apr 2014 - Feb 2018]

  • Designed end to end license plate detection system written in Bengali language from high speed moving vehicles. Applied state of the art deep learning and computer vision algorithms, Python and TensorFlow to develop the system.

  • Developed machine learning algorithm to measure image sharpness and the presence of eyeglass in the face image. These algorithms are used as a part of face verification system, standardized by ICAO. Used scikit-learn and C++ to implement proposed algorithms.

  • Contributed in face verification project using numerical analysis to reduce image template size by 75%. Applied C language to use the algorithm in embedded systems.

  • Proposed an algorithm to improve a speaker recognition algorithm. Used deep learning based autoencoder technique to improve the speaker search time and accuracy.

  • Developed algorithm to search palm print image from large database. Used computer vision techniques and machine learning algorithms (K-Means clustering, Gaussian Mixture Model) to design the search algorithm.

  • Gained first-hand experience to design and develop deep learning and machine learning projects for industrial application. Used computer vision and numerical analysis to optimize the end product.


KAZ Software Ltd, Software Engineer

[Oct 2013 - Apr 2014]

  • Designed a web-based system to aggregate information from multiple websites using user interaction. Used .NET, and SQL server to design the backend system.

  • Developed interactive browser extension as backend interface. Used JavaScript frameworks to implement the extension.

  • Designed end to end interactive backend and frontend web solution using state of the art frameworks.


Nascenia IT Ltd, Software Engineer

[Jun 2012 - Oct 2013]

  • Used Ruby on Rails web framework and object relation PostgreSQL database to design and implement a home appliance controller system (LockState Connect). Developed API end point for third party users.

  • Learnt how to interact with the client, design business idea, set and achieve deadline

Developer Profile

Kaggle

I regularly take part in various data science competitions hosted at Kaggle. This is the largest community of data scientists in the world. Currently, I am ranked as the top 3% data scientist among the 100K registered users. Before starting my Ph.D., I reached into the top 150 position in multiple data science competitions.

Github

  • Used Ruby on Rails web framework, to implement a user validation system using captcha for open source and distributed social network site, Diaspora.

  • Privly is an open-source content privacy system. It allows a user to post content on any website, keeping full control over the shared post. Integrated a payment method and secured token sharing system.

  • Implemented k-best feature functionality into the SelectFromModel method for scikit-learn library. It enables users to find the topmost important features from a machine learning model.

  • Contributed to implementing a case insensitive URL difference checker for a Google project named url_diff.

Stackoverflow

Competitive Programming

  • Solved over 300 problems in UVa online judge platform using the C++.

  • Achieved silver level for both Python and problem-solving track in HackerRank competition platform.

Upwork

https://www.upwork.com/o/profiles/users/~0149cab5ba9acaf3ba/

  • Worked as a freelance software developer at Upwork. Successfully completed 25 jobs using PHP, Python, C# and MySQL.