Dr. Amir Dehsarvi

Dr. Amir Dehsarvi

I am a researcher interested in the application of machine learning techniques for the analysis of medical data.

I am a Machine Learning Engineer & Data Scientist with expertise in AI-driven medical technologies, neuroimaging, and high-performance computing. I have extensive experience developing and optimizing ML/DL models for healthcare applications, with a strong background in Python, MATLAB, R, and deep learning frameworks. My work demonstrates a proven ability to build scalable data pipelines, automate complex workflows, and integrate AI solutions into real-world applications. I am passionate about leveraging data science and ML to drive innovation in health tech and beyond.

Professional Experience

  • 2022 to Date: Postdoctoral Researcher, The Institute for Stroke and Dementia Research (ISD), Ludwig-Maximilians-Universität München (LMU), University of Munich (Germany).
    • Developed and integrated an automated neuroimaging preprocessing pipeline (using Python, MATLAB, R, Shell) for multiple modalities (MRI, fMRI, PET, DTI) into the GRIP platform, a Gates Ventures initiative, optimizing data quality and analysis. This pipeline’s effectiveness is demonstrated by its use in over 25 peer-reviewed publications (2023-present) in neurodegenerative disease research.
    • Developed, validated, and deployed a deep learning model to infer full Alzheimer’s disease A/T/N classification from single tau-PET scans, achieving high predictive accuracy for amyloid-PET (r=0.8) and MRI grey matter density (r=0.76).
    • In addition to my research, I manage the High-Performance Computing (HPC) resources and provide support to research labs for their data analysis needs on the HPC.
  • 2018 to Date: Chief Technology Officer, ClearSky Medical Diagnostics Ltd., York (UK).
    • Contributed to the development of machine learning (ML)-based medical devices for diagnosing and monitoring neurodegenerative conditions, including PD-Monitor, LID-Monitor, and MCI-Monitor.
    • Optimized ML for movement disorder analysis, enhancing diagnostic precision.
    • Collaborated with multidisciplinary teams, including clinicians and engineers, to integrate AI-driven solutions into clinical applications.
  • 2021-2022: Postdoctoral Research Associate, University of York (UK).
    • Applied white-box machine learning to resting-state fMRI data to differentiate depression from healthy controls.
  • 2020-2021: Research Fellow, University of Aberdeen (UK).
    • Used neuroimaging to define a fatigue-related brain network in rheumatoid arthritis, exploring how therapies impact it for potential DBS or similar targeting.
  • 2021: Machine Learning and Image Processing Engineer, smartR.ai, Edinburgh (UK).
    • Created deep learning pipelines to normalize and match color profiles between FIBI and H&E histological images.
  • 2019: Post-Doctoral Researcher in Neuroimaging, The University of Dublin (Ireland).
    • Linked speech patterns to brain volume changes in MCI/AD, exploring speech as an early marker of cognitive decline.
  • 2016-2017: Professional Engineer, My Therapy Tools Ltd.
    • Provided professional engineering support to a Horizon 2020 telerehabilitation platform development for brain injury patients.

Research Focus

My research focuses on the application of machine learning/deep learning for analyzing biomedical data, including brain imaging, movement, and speech, for disease diagnosis and target identification. I lead the development of novel end-to-end analyses, particularly in diseases such as Parkinson’s disease, Alzheimer’s disease, autism, and depression. At the LMU Hospital, I am working with Dr. Franzmeier on developing cutting-edge pre-processing and processing pipelines for multimodal brain imaging analysis (MRI, fMRI, PET, DTI, EEG, etc.).

Collaborations

I have collaborated with renowned researchers such as Prof. Smith, Dr. Waiter, Prof. Basu, and Prof. Reilly on multiple projects focusing on objective assessment of depression from rsfMRI brain scans, investigating underlying mechanisms of rheumatoid arthritis-related fatigue in the brain, and further, contributed to projects analyzing speech and brain imaging features for the classification of Alzheimer’s disease patients and mild cognitive impairment.

Education

  • 2014-2018: PhD Electronic Engineering, University of York (UK).
    • Supervision: Prof. Smith.
    • Thesis: Cartesian Genetic Programming Classification of Resting-State fMRI: Towards a Brain Imaging Biomarker for Parkinson’s Disease.
  • 2013: MSc Digital Signal Processing, University of York (UK).
  • 2010: BSc Applied Science Electronics, University of Science and Arts of Yazd (Iran).

Technical Skills

Machine Learning, Deep Learning, Git, Neuroimaging data analysis, High Performance Computing, MATLAB, C, R, Python, Shell, PHP, SQL, Docker, Linux, AI-based medical devices.