Manuel Jahn
PHD Candidate
Basel, Basel-City, Switzerland
Summary
🤩
Rockstar🎓
Top SchoolManuel Jahn is a PhD candidate at the University of Basel with six years of machine learning and data science experience spanning academia, startups, and consulting. He focuses on improving interpretability in medical AI, aiming to translate research into better cardiovascular disease prediction in clinical settings. His research pedigree includes a best-paper award at CVPR 2021 for work on transformers in image generation, and practical ML engineering contributions to the well-known CompVis/taming-transformers repository, where he improved data pipelines and dataset handling. In industry he developed and deployed causal ML models for marketing at TNG and led data science efforts at Geoblink, blending technical depth with team leadership. Trained in physics and applied computer science at Heidelberg, he brings strong quantitative foundations and multidisciplinary problem-solving skills. Fluent in German, English, and Spanish (with strong Italian and basic Portuguese/French), he is comfortable working across international, cross-functional teams.
6 years of coding experience
Master's degree, Physic, Master's degree, Physic at HeidelbergBachelor of Science - BS, Applied Computer Science, Bachelor of Science - BS, Applied Computer Science at Heidelberg University
Github Skills (10)
data-preprocessing10
computer-vision10
pytorch10
machine-learning10
data-pipeline10
data-pipelines10
python10
datasets10
data-set10
transformers9
Programming languages (2)
Github contributions (5)
CompVis/taming-transformers
Aug 2021 - Jan 2022
Taming Transformers for High-Resolution Image Synthesis
Role in this project:
ML Engineer
ML EngineerContributions:3 reviews, 16 commits, 2 PRs in 4 months
Contributions summary:Manuel made significant contributions to data loading and processing pipelines within the repository, specifically for datasets like AnnotatedObjectsCOCO and Open Images. They implemented a scene image sampler and a mechanism to handle class compatibility, indicating a focus on model training and data preparation. Further contributions included modifications to the dataset structure and utilities, suggesting a role in refining the data ingestion and preprocessing components.
pytorchtransformersimage-synthesisdeep-learningsynthesis
manolo-lolo/potts-model-monte-carlo
Dec 2019 - Dec 2019
Contributions:1 commit, 14 pushes, 1 branch in 1 day
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.