Simon Kohl
Founder & CEO at Latent Labs
London, England, United Kingdom
Summary
👤
Senior🎓
Top SchoolSimon Kohl is a founder and CEO with nine years of deep technical and research experience at the intersection of machine learning, computational biology and medical imaging. He was a core member of DeepMind’s AlphaFold2 team and later led protein design efforts and wet-lab setup at the Francis Crick Institute, blending algorithmic expertise with laboratory implementation. His PhD work at DKFZ produced generative models for ambiguous medical image segmentation, reflecting a focus on uncertainty-aware AI for clinical use. An active open-source contributor, he improved data-augmentation tooling for 2D/3D image pipelines (notably enhancements to MIC-DKFZ/batchgenerators) that support robust training in biomedical tasks. He also co-founded the Heidelberg AI meetup and organized interdisciplinary events on AI and art, demonstrating a knack for community building and translating research to broader audiences. Based in London, Simon combines rigorous academic training in physics and computer science with practical startup leadership and hands-on ML engineering.
9 years of coding experience
7 years of employment as a software developer
Physics, Physics at Universidad de AntioquiaDoctor of Philosophy (PhD) Computer Science, Doctor of Philosophy (PhD) Computer Science at Karlsruhe Institute of Technology (KIT)Physics, Physics at Oregon State UniversityGerman Abitur, German Abitur at Hartmanni-Gymnasium Eppingen
English, German, French, Spanish
Github Skills (8)
unit-testing10
data-augmentation10
python10
image-processing10
numpy10
scikit-image8
computer-vision8
machine-learning7
Programming languages (4)
Github contributions (5)
MIC-DKFZ/batchgenerators
Mar 2017 - May 2017
A framework for data augmentation for 2D and 3D image classification and segmentation
Role in this project:
ML Engineer
ML EngineerContributions:9 commits in 1 month
Contributions summary:Simon primarily contributed to data augmentation techniques within the `batchgenerators` repository, a framework for image classification and segmentation. Their work included adding features to mirror data along specific axes, enabling random cropping with margins, and implementing a generator to simulate rigid registration errors. They also added tests for the axis mirroring functionality and an invertible rotation generator, demonstrating a focus on enhancing the framework's data preprocessing capabilities.
classificationdata-augmentationdeep-learningcomputer-vision3d-image
SimonKohl/Lasagne
Mar 2017 - Jan 2023
Lightweight library to build and train neural networks in Theano
Contributions:15 pushes, 1 branch in 6 years
deep-learningtheanoneural-networksmachine-learningneural-network
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.