Philipp Arras

Portrait

Disclaimer: All content on this website reflects my personal views, work, and interests only. It is entirely independent of my employment and does not represent the views, positions, or activities of my employer (MongoDB, Inc.).

Intro

I work as a Pre-Sales Solution Architect for MongoDB.

Previously, I focused on medical imaging and co-founded LambdaFields GmbH. Between 2022 and 2024, I was employed by Bain & Company as a strategy consultant.

If you would like to discuss anything related to my previous science projects (e.g., resolve, nifty or information field theory in general), the life of a physicist at a MBB consulting company, the application of Bayesian imaging algorithms to medical imaging, or how to found a startup, feel free to reach out to me.

Previous Academic Work

I work on Bayesian medical imaging, currently with focus on Magnetic Resonance (MR) imaging, leveraging my imaging tools from astrophysics.

While seemingly distinct, astronomy and medical imaging share fundamental principles. Just as radio telescopes measure data in the Fourier domain (or k-space), so does MR imaging. Furthermore, astronomical techniques like Galactic tomography resemble medical imaging modalities such as CT and PET scans. Recognizing these strong parallels, I am dedicated to applying my advanced Bayesian imaging algorithms, originally developed for radio interferometry in astrophysics, to the field of medical imaging. This allows to leverage powerful inference techniques to extract maximum information, achieve maximum resolution, and provide robust uncertainty quantification for medical imaging.

I did my PhD and a postdoc at the intersection of computational statistics and physics. At the Max-Planck Institute for Astrophysics, I developed Bayesian statistics algorithms for imaging radio telescope observations. My aim has been to squeeze out all the information present in radio data sets in order to obtain the highest possible resolution together with uncertainty quantification. To this end, I developed theoretical models for inference algorithms and tailored them for unification of calibration and imaging, multi-spectral imaging, polarization imaging, data fusion with single dish data, and Very Long Baseline Interferometry (VLBI).

These ideas are condensed into various software projects: NIFTy (a Python library for Bayesian inference of fields), resolve (my radio imaging library), vlbi_resolve (the same thing for dynamic VLBI imaging) and ducc (which includes a non-uniform FFT specialized for radio interferometry). I hold a PhD in physics and was part of Rüdiger Westermann’s group (TU Munich) and Torsten Enßlin’s information field theory group (MPA Garching).

During my masters at Heidelberg University I worked on String Theory, Quantum Field Theory and pure maths.

Papers and Publications

Papers

See also: arXiv, Google Scholar

Theses

Misc

Software

Talks

Invited Talks

Contributed Conference Talks

Workshops (co-)organized by me

Attended workshops

Teaching

Co-supervised master students

More resources

Misc

Contact

Impressum, Datenschutzerklärung, Haftungsausschluss