Research Intern – Deep Learning & Remote Sensing

Developed a flood detection model using SAR satellite data, enhancing early warning systems with a 92% improvement in detection accuracy. Utilized Google Earth Engine (GEE) and deep learning techniques for real-time land cover classification. Worked with MODIS, Landsat, and Sentinel data to analyze and map flood-prone regions.

Machine Learning Researcher

Built a semi-supervised CNN-PINN model for infrastructure crack detection, improving accuracy by 85%. Integrated FEA stress simulations with AI to generate Grad-CAM and SHAP-based explainability maps. Focused on interpretable and physics-informed AI for real-world structural health monitoring.

Machine Learning Lead

Led ML/AI initiatives and technical workshops for student developers. Mentored peers and promoted collaborative learning in deep learning and applied AI projects. Guided student teams in hackathons and research competitions.