Md. Ittesaf Hossain (IH Arik)
Final-year CSE student, applied AI researcher, and full-stack systems engineer.
I operate at the intersection of applied deep learning research and modern full-stack engineering. My approach focuses on building robust, type-safe applications (Next.js/TypeScript) backed by modular, high-throughput asynchronous services (FastAPI/Python).
As a researcher at the Visual Data Analysis Lab (VDAL), Green University of Bangladesh, my academic focus targets geospatial bitemporal imagery analysis. My thesis project investigates how bi-temporal change-detection models can compare pre- and post-disaster satellite captures to automate damage assessment, replacing slow and manual mapping procedures.
I also collaborate with small teams and startups to build SaaS products — from an AI business communication platform to a study-abroad information portal — with the goal of writing clean, maintainable systems that ship real-world value.
Timeline
Graduate Researcher & SaaS Engineer
Developing transformer-based bi-temporal change detection models for satellite damage assessment. Engineering Next.js/FastAPI applications for client operations.
Lead author — two IEEE conference papers
Explainable stacking ensembles (Random Forest, CatBoost, LightGBM, XGBoost) with SHAP and LIME for thyroid disorder detection and diabetes risk prediction. Co-author on an explainable kidney stone detection paper using a dynamically weighted CNN ensemble with Grad-CAM++.
Full-Stack SaaS Systems Lead
Built an AI business communication platform (GoCustify AI / Mabdel), a fantasy basketball app (Mon5majeur), and a study-abroad information platform (MyFutureAbroad).
Applied Computer Vision Architect
Built YOLO-based PPE detection apps for construction-site safety and a ResNet18 brain tumor MRI classifier with a RAG knowledge base.
B.Sc. in Computer Science & Engineering (Final-Year)
Focusing coursework on core neural networks, computer vision, database optimization, and software architecture.