Malte Londschien

Malte Londschien

ETH AI Center Doctoral Fellow

Software

  • rgbm: A lightweight, Rust-native gradient boosting machine. Designed as an alternative to LightGBM, rgbm focuses on core functionality, prioritizing code readability and maintainability.
  • ivmodels: Instrumental variables regression in Python. Software package accompanying Londschien and Bühlmann (2024). See docs. Available on PyPI and conda-forge.
  • anchorboosting: Efficient tree-based nonlinear anchor regression and classification for Python.
  • icu-features: Simple feature engineering and transparent outcome definitions for ICU data.
  • changeforest: Classifier-based change point detection. Software package accompanying Londschien, Bühlmann, and Kovács (2023). Written in Rust, with R and Python bindings. Available on crates.io, PyPI, and conda-forge (R, Python).
  • biosphere: Simple, fast random forests in Rust. Dependency of changeforest. With Python bindings. Available on crates.io, PyPI, and conda-forge.
  • crispybeefbot: Get notified about your favorite meals at the ETH Mensa.

I love open-sourve and have made contributions to open-source software such as dask, LightGBM, numpy, pandas, polars, pytorch-lightning, scikit-learn, and scipy.