Welcome to Tanviranik's Homepage

I am a passionate developer and researcher having sincere interest in data engineering, analytics, and machine learning.

About Me

I am Tanvir, graduated from the department of Computer Science and Engineering, Bangladesh University of Engineering and Technology in 2012. I have around 8 years of industry experience in working with business intelligence. I have worked for IQVIA, Robi Axiata Limited, and ACI Limited within last 8 years of my career.

Please download my resume from this link. Visit my academic summary page. Have a look at my Publications.

My Profile

My profile at LinkedIn, where I maintain my professional networks.

My profile at Google Scholar, where my research and publication links are available.

Few research works on HCI department at BUET.

My profile at DBLP , listing a few of my research works.

My profile at Github , containing some repositories and notebooks for machine learning.

My profile at Academia.

Recent Projects

Credit Customer Analytics

We performed full suit credit customer analytics. This included segmenting credit customers using unsupervised machine learning techniques. Besides, we detected fraudulent cusomers with a fraud detection model. Key features:

  • Customer clustering
  • Fraud Detection
  • Credit Collection Forecasting

Custom Object Detection

Using YOLO V3, we developed custom models able to detect custom objects from images. This project was done for a merchandizing brand. They have around 10 products for which they want to develop a machine learning model capable of detecting their products from images collected by field forces through a mobile app. Key features:

  • YOLO V3
  • OpenCV
  • GPU accelerated

Real-Estate Pricing Analysis

A web application for geo-spatial analysis that compares real-estate property prices collected through web scraping from different websites. The pricing model is used to take decision whether to buy a property or not.

  • Geo-Spatial Analytics
  • Google Map / Mapbox MAP
  • Data scraping

Forex Trading Forecasting

This web application is backed by Azure AutoML machine learning. An AutoML script is running every hour with last 1 year up-to-date forex trading dataset to train a model. This model is used to predict forex EURO-To-USD convertion rate for next 5 minutes. Key features:

  • Azure AutoML
  • Machine Learning