Tasfiq Kamran

about me

السَّلَامُ عَلَيْكُمْ (May peace be upon you)

Hi, This is Md. Tasfiq Kamran born and raised in Dhaka, Bangladesh. I'm passionate about Data Science, AI and Machine Learning. Interested in topics like Computer Vision and NLP tasks. Currently I'm working as a TA (Data Science) at MasterCourse BD,.. Prior that I worked as a Machine Learing Engineer for 6 months at Innovative Skills BD.

I enjoy working with data. I find it quite insightful and fun when exploring and visualizing data. As if the data have something to say. One of my personal project where I worked to create a prediction model on the property prices of Bangladesh, which I've found quite interesting while I was exploring the data. You can check out the project here. Also checkout my other projects on the Projects section. I have obtained skills so that I can work with data and provide data driven solution and also to put an impact to the society. I love challenges and keen on learning.

Research Interests:

• Computer Vision on Biomedical Topics
• Using AI to solve social problems.
• NLP

Education:

B.Sc (Engineering) in Electrical and Electronic Engineering (2017 - 2021)
Hajee Mohammad Danesh Science and Technolgoy University Dinajpur, Bangladesh

Skills

Programming Languages & Databases

Data analysis & Visualization

ML & DL Frameworks

Web Scraping & Development

Development Tools

Experience

Omdena Inc. USA

Junior Machine Learning Engineer (October 2024 - Present)

MasterCourse BD,. Bangladesh

TA (Data Science) (June 2024 - Present)

Innovative Skills BD,. Bangladesh

Machine Learning Engineer & TA (Deep Learning with Computer Vision) (December 2023 - May 2024)

projects

Who Are You!

Period: February 2024 - April 2024 Motivation: A deep learning based Face Recognition system, which can be easily integrated in attendance system, security system etc. I worked on this project while I was working in Innovative skills BD as an Machine Learning Engineer. The idea was to develop a depp learning based Face Recognition system that can be adapted to various situations. Outcome: I worked on the development of the DL based system as well as software development with the software development team. Contact me for more info.

Just Transcribe and Summarize

Period: September 2023 - Octorber 2023 Motivation: This was the final project under the supervision of Mohammad Sabik Irbaz at MasterCourse. The idea was to create deploy a project which can summarize, classify and extracts important topics from the given video or audio file/link or long plain texts. What motivated me to this project was to, in a hurry situation and you want to understand and get to know the important topics of lecture video or news video or a long paragraph, this will certainly come handy. Outcome: Eventually I could pull of this. I've used the medium article data from kaggle for traning the classification model. More about the project is on my github repo. Check that out if you feel interested. Project on Github Deployment

From Paper With Tasks

Period: August 2023 - September 2023 Short Description: A multilable text classification project which predicts the tasks related to the given abstract of a research paper. This was the third project under the supervision of Mohammad Sabik Irbaz at MasterCourse. Motivation: As I was going through the paperswithcode.com (where I colldected the data from) and reading the different abstracts of diiferent papers and eyeing the task section of that paper, I was struck with an idea that, what if I buil a classification model, where I will put the abstract of a paper and it will output the task that can be done using the idea of that paper, or sort of a recommendation system where I will be able to know what can be done using the idea of the paper. Going with this idea, I started to collect data from the paperswithcode.com through scraping and built my own From Paper With Tasks. Inspired by the movie named From Paris With Love. Checkout the github repo to learn more Outcome: Built the multi-label text classification model and deployed on onrender.com Project on Github Deployment

Amazon Data Science Book Analysis

Period: June 2023 - July 2023 Short Description: A data analysis project, where I utilize python to scrape data from amazon and Tableau software to analyze and visualized the data. Motivation: I like books on Data Science and Machine Learning. So, for a data analysis project I thought why not scrape some data science/ML related books from amazon. I will be able to get some insights about which publisher publishes more DS books, their ratings, price etc. With that in mind, started scraping with selenium, and analyzed and visualized the data using Tableau. Outcome: Colltected about 1300+ books from amazon and created a csv file, preprocessed the data using pandas and analyzed using Tableau. Alos published the report on Tableau. Project on Github Deployment

Chicken Breeds Recognizer

Period: May 2023 - June 2023 Short Description: An image classification project. As the name suggests, this project classify chicken breeds from image. Motivation: This was the first project under the supervision of Mohammad Sabik Irbaz at MasterCourse. I was asked to do an image classification project, but, seriously I was quite stuck about the project idea. Then came the chicken xD. However, I found that there are a lot of chicken breeds out in the world, so, I thought why not do a project that classifies chicken. So started gathering image data and collected enough data to classifiy 20 chicken breeds. Yes, the trained model can classify 20 chicken breeds. More here. Outcome: It didn't took too much time to pull up this project. Collected the data, preprocessed, trained the model using fastai again preprocessed. Eventually I got what I needed. Finally, deployed on github io and huggingface Project on Github Deployment Huggingface App

BD Property Price Prediction

Period: March 2023 - May 2023 Motivation: I really enjoyed doing this project. As the data is collected from Bangladesh and as a Bangladeshi I enjoyed working on this project The dataset is comprised of different property information from two major cities (Dhaka, Chattogram) of Bangladesh. You can find more info here. After thorough analysis of the data and lots of preprocessing, finally I used some of the classical regression ML models to train. Which gave me quite satisfactory result despite the data being imbalanced. Finally, I used streamlit to build a web app where you can get predicted price based on the provided information like location, # of beds, # of baths etc. How cool is that! Outcome: Took a little while to do this project. As analyzing and preprocessing the data took a long time. Check out the github repo if you feel interested. Project on Github Deployment

Resume/cv

Tasfiq Kamran (CV)

Contact

• Twitter
• Facebook
• Instagram
• GitHub

Elements

i = 0;

while (!deck.isInOrder()) {
    print 'Iteration ' + i;
    deck.shuffle();
    i++;
}

print 'It took ' + i + ' iterations to sort the deck.';