Malte Hoffmann
Assistant Professor Of Radiology at Massachusetts General Hospital
Boston, Massachusetts, United States
Summary
🤩
Rockstar🎓
Top SchoolMalte Hoffmann is an Assistant Professor of Radiology at Harvard Medical School and research staff at Massachusetts General Hospital with eight years of experience at the intersection of MRI physics and medical image machine learning. He earned a PhD in MRI Physics from Cambridge and applies that deep signal-processing background to develop unsupervised image-registration and synthesis methods, contributing code improvements to the widely used VoxelMorph project by enhancing its SynthSeg synthesis pipeline. His work bridges rigorous academic research and practical software engineering—refactoring generative models, stabilizing loss terms, and improving training and model-saving workflows for reproducible experiments. Affiliated with HST and Boston Children's Hospital, he combines clinical collaboration with open-source impact, enabling translational tools for radiology. Colleagues describe him as a hands-on scientist-engineer who turns complex MRI physics into usable ML tooling.
8 years of coding experience
Doctor of Philosophy (PhD), MRI Physics, Doctor of Philosophy (PhD), MRI Physics at University of CambridgeSelective Bachelor's degree (Licence et Magistère 1), Fundamental Physics, Selective Bachelor's degree (Licence et Magistère 1), Fundamental Physics at Paris-Sud University (Paris XI)
German, French, English
Github Skills (16)
algorithm10
itk10
data-structures10
algorithms10
machine-learning10
unsupervised-learning10
c-language10
deep-learning10
tensorflow10
cprogramming-language10
surf10
image-registration10
data-structure10
neuroimaging10
surface10
Programming languages (3)
Github contributions (5)
voxelmorph/voxelmorph
Apr 2020 - Oct 2022
Unsupervised Learning for Image Registration
Role in this project:
ML Engineer
ML EngineerContributions:45 commits, 68 PRs, 6 pushes in 2 years 6 months
Contributions summary:Malte primarily focused on enhancing the SynthSeg model for image synthesis within the unsupervised learning framework. Their commits demonstrate modifications to the SynthSeg model, including training on synthesized image-segmentation pairs, fixing loss terms, and correcting model saving procedures. Furthermore, the user refactored the codebase, moved the generative model to Neurite and implemented smoothing and normalization techniques.
pytorchunsupervised-learningdeep-learningunsupervisedoptical-flow
mu40/freesurfer
Dec 2017 - Dec 2024
The official Freesurfer repository maintained by LCN
Contributions:290 pushes, 147 branches in 7 years 1 month
freesurferlcn
Find and Hire Top DevelopersWe’ve analyzed the programming source code of over 60 million software developers on GitHub and scored them by 50,000 skills. Sign-up on Prog,AI to search for software developers.