AHMAD SAMANDARI

Developed a framework integrating XAI and vision-based monitoring to enhance object detection in autonomous driving systems. Utilized rule-based decision boundaries and cumulative feature ranking to ensure transparency and adaptability, addressing domain shifts and environmental variability. Enabled adaptive preprocessing and interpretable monitoring to improve system performance and safety under dynamic real-world conditions.

Developed and fine-tuned object detection models using YOLOv5, YOLOv8, and DETECTRON2 to optimize detection accuracy and efficiency in diverse settings. Conducted extensive model tuning and analysis to achieve high performance, leveraging state-of-the-art architectures for object detection in challenging environments, with particular focus on applications within autonomous systems.

Conducted information extraction from Electronic Health Records (EHRs) to improve healthcare system quality and data interpretability. Designed algorithms to extract and organize critical health information, enhancing EHR interpretability and aiding clinicians in decision-making, while contributing to data-driven quality improvements in healthcare systems.

Researched physiological and mental signals to identify workload patterns in workers, contributing to the development of stress management solutions.