M. Moein Shariatnia

prof_pic_new.jpeg

London, UK

MD and ML researcher specialising in medical imaging, multimodal learning, LLMs, and generative AI. UK Global Talent Visa holder. 350+ citations (h-index 6) and 4+ years of large-scale R&D, with collaborations at Google DeepMind, Mayo Clinic, Stanford, Harvard, and Hospital for Special Surgery. Published at ICML, ICLR, MICCAI workshops and leading medical journals, and co-secured >$70k in competitive grants.

I’m currently collaborating with the AI team at Hospital for Special Surgery (HSS) under Dr. Ayoosh Pareek, building state-of-the-art models for bone age prediction, landmark detection, and foundation models for MSK imaging. Earlier research includes work on contrastive and transfer learning (ICML’22, ICLR’23) with Google DeepMind/UW researchers, and TransDeepLab for medical image segmentation (~200 citations, MICCAI’22).

On the engineering side, I’ve shipped end-to-end ML products including a multimodal trademark search pipeline on Azure and a face-health Computer Vision application with Django REST + Ray Serve. I also maintain open-source tools: Simple CLIP (700+ GitHub stars), CONFLARE (6k+ PyPI downloads), and LatteReview for multi-agent systematic reviews.

I graduated with my MD from Tehran University of Medical Sciences (top 0.01% national exam, full scholarship) and am based in London, UK.

selected publications

  1. hands.jpg
    Determination of Skeletal Age From Hand Radiographs Using Deep Learning
    Joshua T Bram, Ayoosh Pareek, Samuel A Beber, Ruth H Jones, and 5 more authors
    The American Journal of Sports Medicine, 2025
  2. iclr.png
    The role of pre-training data in transfer learning
    Rahim Entezari, Mitchell Wortsman, Olga Saukh, Moein Shariatnia, and 2 more authors
    MRL Workshop, ICLR 2023, 2023
  3. shoulder-ai.jpg
    Deep learning model for measurement of shoulder critical angle and acromion index on shoulder radiographs
    Moein Shariatnia, Taghi Ramazanian, Joaquin Sanchez-Sotelo, and Hilal Maradit Kremers
    JSES Reviews, Reports, and Techniques, 2022
  4. transdeeplab.png
    TransDeepLab: Convolution-free transformer-based DeepLab v3+ for medical image segmentation
    Reza Azad, Moein Heidari, Moein Shariatnia, Ehsan Khodapanah Aghdam, and 3 more authors
    In PRIME Workshop, MICCAI 2022, 2022