Nils Lehmann

new_headshot2.jpeg

I am currently a PhD Student at the Technical University of Munich (TUM), supervised by Prof. Jonathan Bamber and Prof. Xiaoxiang Zhu, specializing in generative modeling and uncertainty quantification for Earth observation applications. I have experience in processing terabyte-scale satellite datasets, developing foundation models for remote sensing, and creating machine learning-ready Earth system data products. More broadly my interests lie in generative modeling for Earth Observational data and building data-driven systems to tackle hard and relevant EO problems that benefit society. I am also very passionate about open-source software and a maintainer of TorchGeo and author of Lighting-UQ-Box. In a past life, I played NCAA Division II basketball.

news

May 12, 2026 New preprint: No One Knows the State of the Art in Geospatial Foundation Models. I’m grateful to have been part of this collaboration led by Isaac Corley. There is still so much work to do in EO benchmarking to bridge the gap between academic research and real-world applications.
Feb 12, 2026 New preprint of an idea we have been working on: EO-VAE as a multi sensor tokenizer for Earth observation data: arxiv, code
Sep 28, 2025 I feel very fortunate to have the chance to visit Ando Shah and Prof. John Chuang at UC Berkeley - School of Information for a research stay

research I have contributed to

  1. GEO-Bench-2: From Performance to Capability, Rethinking Evaluation in Geospatial AI
    Naomi Simumba*, Nils Lehmann*, Paolo Fraccaro*, and 9 more authors
    Transactions on Machine Learning Research, Jul 2026
    arXiv HTML PDF
  2. No One Knows the State of the Art in Geospatial Foundation Models
    Isaac Corley, Nils Lehmann, Caleb Robinson, and 6 more authors
    arXiv preprint arXiv:2605.12678, Jun 2026
    arXiv
  3. EO-VAE: Towards A Multi-sensor Tokenizer for Earth Observation Data
    Nils Lehmann, Yi Wang, Zhitong Xiong, and 1 more author
    In ICLR 2026 Workshop on Machine Learning for Remote Sensing (ML4RS), Main Track, Apr 2026
    arXiv HTML PDF Code
  4. Panopticon: Advancing Any-Sensor Foundation Models for Earth Observation
    Leonard Waldmann, Ando Shah, Yi Wang, and 6 more authors
    In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) Workshops, Jun 2025
  5. Lightning UQ Box: Uncertainty Quantification for Neural Networks
    Nils Lehmann, Nina Maria Gottschling, Jakob Gawlikowski, and 3 more authors
    Journal of Machine Learning Research, Jun 2025
  6. SSL4EO-L: Datasets and Foundation Models for Landsat Imagery
    Adam Stewart, Nils Lehmann, Isaac Corley, and 6 more authors
    In Advances in Neural Information Processing Systems, Jun 2023