Scientist at Helmholtz Imaging, Max Delbrück Center (MDC), Berlin. I design open-source software that makes scientific image analysis reusable and replicable, and I develop image-analysis methods that include machine learning. Currently I am exploring how LLM agents can orchestrate verified, executable tools while every step stays reproducible.
Projects
- Album: framework for sharing and replicably running scientific routines, each in its own environment. Owner and architect; maintainer of the conda-forge package. Code on GitLab · arXiv:2110.00601
- Polarity-JaM: toolbox for cell polarity, junction and morphology quantification, with a napari plugin and an R/Shiny web app. Software architect and project co-lead. Nature Communications 2025
- ARI3D: interactive quantification of regions in 3D X-ray CT images. Architect and lead developer. arXiv:2508.09849
Album and ARI3D live on GitLab, where most of my development happens: gitlab.com/jpalbrecht91.
I also run the programming exercises and tutorials of the course Machine Learning for Image Analysis at HU Berlin and HPI.