Mufti Ahmad

I’m a final-year Computer Science student with hands-on research experience in machine learning, deep learning, computer vision, explainable AI, and predictive analytics. I’m passionate about building efficient, data-driven AI solutions for real-world applications.

My work includes developing an explainable and scalable framework for compost maturity prediction (ACM Digital Library, 2024), where I engineered physical, chemical, and biological features and evaluated ensemble models with explainability techniques. I also designed a multimodal risk assessment and monitoring framework for assisted elderly living (FYDP), including dataset curation and video annotation for YOLO-based behaviour detection and a real-time alerting layer for caregivers.

Across projects, I’ve built and improved ML/DL pipelines for pneumonia classification for low-compute environments, AQI prediction with SHAP-based feature importance for interpretability, and OCR frameworks for Shadhu Bangla manuscripts using Roboflow-driven dataset versioning, augmentation, and quality control. I aim to combine strong modeling with clear interpretability and careful data workflows.