Highly specialised consulting and research services that bring state-of-the-art AI models — including foundation models — into clinical radiology research, and see them through to publication.
AI is transforming medicine — and radiology in particular — at an unprecedented pace. Yet there is a critical shortage of experts who can bridge complex AI architecture and everyday clinical practice.
Dr. Rahmanzadeh Medical AI exists to close that gap: offering highly specialised consulting and research services that integrate cutting-edge AI models — including foundation models — into clinical research, so clinicians and imaging centres can put modern AI to work in real scientific projects rather than treating it as a black box.
Support for clinicians and radiological centres who want to actively use AI in research projects — from first model to final publication.
Implementation and validation of modern architectures — such as multi-encoder nnU-Net or F3-Net — on real clinical datasets, with methodology built to hold up scientifically.
Training physicians in data pre-processing, GPU inference and model architecture — taught hands-on within shared projects rather than as abstract theory.
Guiding AI projects through to scientific publication, raising the academic visibility of the clinicians and centres behind them.
A combination of clinical and technical qualification that is rare in Switzerland.
A radiologist and AI engineer — combining clinical practice with deep learning research to bridge complex AI and the radiology clinic.
Whether it's implementing a model, training your team, or taking a project through to publication — I'd be glad to hear what you're working on.
✉ ContactA radiologist and scientist working across medicine, machine learning and research — one of the few people fluent on both sides of the scanner.
I'm a radiologist trained in radiology and neuroradiology (MD), and in AI and biomedical engineering (PhD & postdoc). My doctorate at the University of Basel — supported by the MSIF McDonald Fellowship and a Swiss Government Excellence Scholarship — focused on deep learning for lesion detection and segmentation in advanced quantitative MRI, followed by an AI postdoctorate at the University of Bern and years of radiology practice and research at Inselspital Bern.
My published work spans 100+ papers in journals including The Lancet, JAMA and Brain, alongside deep-learning architectures such as RimNet and GAMER-MRI for multi-contrast MRI. I sit on the editorial boards of Frontiers in Neurology and Frontiers in Neuroimaging, and hold 10+ U.S. patents.
I founded UltraMed on this rare pairing of clinical and technical qualification. As a radiologist who has both read the studies and trained the models, I help clinicians and radiological centres apply state-of-the-art AI in real research — from implementation and validation, through hands-on training, to scientific publication.