Mohammadreza Riyazat

Background

About Mohammadreza Riyazat

I’m a PhD student at the AI–Medical Imaging Lab. My passion is building deep-learning pipelines that blend classic computer-vision techniques with modern image-processing to tackle tough biomedical-imaging challenges. Right now I’m working on federated adversarial learning for brain-tumour segmentation. I’m teaching a model to learn from MRI scans stored at several hospitals without moving the data, while also making it resilient to privacy attacks and domain shifts. That pushes me to think hard about secure data engineering, distributed optimisation, and how to prove a model really generalises. Before AI, I spent six years in advanced control engineering, taming nonlinear and multi-agent systems. All that stability theory, fractional calculus, and optimisation shaped the maths toolkit I now lean on to keep today’s data-driven models robust and explainable. I’m drawn to projects that balance solid theory with real-world impact—especially in medical AI, privacy-preserving machine learning, and any work that turns clever mathematics into better healthcare. Outside the work, I unwind with brutally challenging Souls-like (Soulsborne) games—their steep learning curves are the perfect workout for persistence and problem-solving.