Salauddin Tapu

Graduate Research Assistant | Gachon University

About Me

Hello! I am Salauddin Tapu, an AI enthusiast with expertise in machine learning, deep learning, and computer vision. The timeless question, "Can machines think?", has sparked my curiosity about the transformative capabilities of AI, which gradually evolved into a focused passion for building intelligent systems that can make a real-world impact. During my undergraduate years, this passion led me to explore how AI can contribute to humankind, and since then, I’ve been fortunate to work on projects that bring cutting-edge research to life. Over the years, I’ve had the privilege of working on a range of impactful, real-world AI projects—building robust systems that bridge theory and application. My work spans from intelligent document analysis to real-time facial recognition, always driven by a desire to solve meaningful problems. Fueled by an insatiable hunger for problem-solving and a passion for intelligent systems, I aspire to tackle real-world challenges through cutting-edge AI research that creates a meaningful, positive impact for humanity.

Academic Background

Gachon University

Mar 2026 - Present | Master's | Chemical and Biological Engineering (Supervisor: Jongkuk Ko)

Khulna University

Jan 2018 - Mar 2023 | B.Sc | Electronics & Communication Engineering Discipline (Supervisor: Dr. Uzzal Biswas)

Research Experience

Graduate Research Assistant

Nov 2023 - Present | ECMS Lab

Work Experience

Anwar Technologies

Nov 2023 - Present | Junior Machine Learning Engineer (Full Time)

Aug 2023 - Nov 2023 | Machine Learning Engineer (Intern)

VISIE Limited

Jan 2023 - Dec 2024 | Junior Machine Learning Engineer (Project-Based)

Projects

Bangla Handwritten OCR

Bangla Text Recognition

Med-2D SegNet

Spider plot comparing the performance of the proposed method with state-of-the-art works, highlighting its superior and comparable Dice score and exceptional performance across various evaluation metrics.

Comparison of model parameters vs. DSC on the KVASIR-SEG dataset. Our proposed method achieved a 95.78% Dice score, outperforming competing models with the lowest parameter count, demonstrating an optimal balance between efficiency and segmentation accuracy.