Kausar Ahmed

Machine Learning Engineer & Researcher || CSE@ULAB

About Me

I'm a passionate Machine Learning Engineer with a strong foundations in artificial intelligence and deep learning. My research focuses on developing innovative solutions for real-world problems using cutting-edge machine learning techniques. I have hands-on experience in computer vision, natural language processing, and explainable AI systems.

I'm particularly interested in Natural Language Processing (NLP) approaches and creating AI systems that are not only accurate but also interpretable and trustworthy. My work aims to bridge the gap between theoretical research and practical applications.

Download CV

Publications


ResNet50 for Glaucoma Detection: Hyperparameter Tuned Transfer Learning on Retinal OCT Images

July 2025 16th International Conference on Computing, Communication and Networking Technologies (ICCCNT) at IIT - Indore, Madhya Pradesh, India

• Developed automated glaucoma detection system using transfer learning
• Achieved 97.62% specificity and 95.35% precision
• Demonstrated impact of selective hyperparameter optimization

Reducing False Alarms in Fire Detection Systems with YOLOv11 and Multi-Sensor Validation

Feb 2025 - [Under Review] Iranian Journal of Electrical and Electronic Engineering (IJEEE)

• Proposed a novel three-layer verification system for indoor fire detection
• Achieved high precision of 0.979 and recall of 0.971
• Integrated YOLOv11 with multi-sensor validation for enhanced accuracy


Research Interests


Machine Learning

Traditional State of the art algorithms and deep learning techniques for solving complex real-world problems.

Natural Language Processing

Focusing on low resource languages (e.g. Bangla) and developing NLP solutions for underrepresented languages.

RAG (Retrieval Augmented Generation)

Multimodal RAG systems that combine retrieval and generation for enhanced AI capabilities.


Honors & Awards


Honorable Mention @ Blockchain Olympiad Bangladesh (BCOLBD) 2025

AI Category · Sept 2025

Officially nominated to represent Bangladesh at the International Data Science Olympiad (IDSOL 2025) in Hong Kong. Awarded an Honorable Mention (Top 6 out of 20 finalists) in the Artificial Intelligence category for our project, "EXIREN: A Multimodal AI-Powered Fire Detection System."

2nd Place @ ULAB CSE FEST Hackathon

July 2025

This hackathon was organized by the Department of Computer Science and Engineering at ULAB as a part of ULAB CSE Fest 2025. The half-day event featured 21 teams competing in a rapid sprint through idea formulation, poster design, prototyping, and presentation—all within a few intense hours. Our team built a web-based recycling solution aimed at promoting sustainable practices on campus.

2nd Runners-up @ National Business Olympiad (NBO) 2022

Higher Secondary Level · Feb 2022

The National Business Olympiad is a platform for young minds to understand what the term "corporate world" really entails. It started journey in March 2021. NBO team members and coordinators are doing their best across Bangladesh. NBO is open to students from schools, colleges and universities across the country. NBO strives to encourage business and other industry students to develop business and personal leadership.


Projects


ScheduLumos

Web application designed to simulate and visualize CPU scheduling algorithms. Particularly useful for students and educators in computer science to learn algorithm behavior. Features interactive visualizations and step-by-step algorithm execution.

Frontend: HTML5, CSS3, JavaScript
Backend: Python, Flask

Live Demo GitHub

GenRoutine

Routine Generator web application designed to assist users in creating and managing personalized routines for academic purposes with intelligent scheduling algorithms. Features automated conflict resolution and optimal time slot allocation.

Frontend: HTML5, CSS3, JavaScript
Backend: Python, Flask

Live Demo GitHub

KlashCard

Versatile flashcard application for vocabulary practice. Features English to Bangla translation but can be easily modified for any learning flashcard type with adaptive learning algorithms. Includes progress tracking and spaced repetition techniques.

Frontend: CustomTkinter (Python)
Backend: Python

Download App GitHub

Education


University

University of Liberal Arts Bangladesh (ULAB)

BSc in Computer Science and Engineering (CSE)
Oct 2022 - Present
CGPA: 3.88 out of 4.00

College

Cantonment College Jashore

HSC (Science)
Jul 2018 - Sept 2020
Result: 5.00 out of 5.00


Contact


Say hello

© KAUS4R   2025