Earlier Software and Machine Learning Projects
Selected software, data, and machine-learning systems built before my Ph.D.
Before starting my Ph.D., I worked on a range of software, data, and machine-learning systems across financial technology, enterprise analytics, e-commerce, NLP, and backend services. These projects helped build my foundation in full-stack engineering, data pipelines, applied machine learning, and production-oriented software development.
Selected Projects
Automated ChatBot using RASA
Developed at Binate Solutions Ltd. from June 2020 to October 2020. Built a conversational chatbot for FAQs, questionnaires, project booking, and appointment scheduling with sales representatives. Used technologies including RASA, React.js, and Docker Swarm.
Warehouse Management System
Developed at ACI Logistics Limited from October 2017 to February 2018. Built workflows for raw materials, processed materials, purchase-order materials, invoices, and inventory reporting for factories and plants.
ChatBot with Speech
Developed at ACI Logistics Limited from November 2017 to December 2017. Built a web-based chatbot using Django, MSSQL, and TensorFlow, with a sequence-to-sequence neural response model trained on internal organizational data.
Sentiment Analysis for Social Media and News
Developed at ACI Logistics Limited from September 2017 to November 2017. Collected Facebook posts, Twitter posts, comments, and newspaper articles through crawling, then classified text sentiment using Google NLP API, IBM Watson NLU, Azure NLP API, Stanford CoreNLP, TextBlob, and VADER.
Web Crawling
Developed at ACI Logistics Limited from July 2017 to August 2017. Built a keyword-based crawler using Django and Scrapy to extract relevant links from target websites with depth control.
Customer Fraud Detection using Neural Networks
Developed at ACI Logistics Limited from July 2017 to August 2017. Built a machine-learning web application for customer fraud detection using Django, jQuery, Bootstrap, MSSQL, and TensorFlow.
Inventory Management System
Built a desktop inventory system using WPF, C#, .NET Framework, MVVM, Entity Framework, Crystal Reports, and MSSQL for branch-level item tracking and rate comparison.
Rule-Based Stemming of Bengali Words
Built a Bengali stemming pipeline using GATE Developer and Snowball scripting, including corpus processing, stopword handling, stemming rules, and integration into a GATE processing workflow.
Image Steganography
Built an application using least significant bit encoding to hide text or image data inside another image.
Other Systems
Additional projects include an ETL data transformation workflow, e-commerce systems, a fault tracking system, Motors Insight, stored procedure execution service, vehicle rent system, financial data portal, digital commercial paper system, and digital building safety management system.
Keywords
Full-stack development, applied machine learning, NLP, enterprise systems, Django, TensorFlow, React, MSSQL, Oracle, Docker, RASA, Scrapy.