Janosh Riebesell - CV

Computational materials scientist working on machine-learning interatomic potentials, high-throughput electronic-structure calculations, and open-source atomistic modeling and analysis software.

  Experience

• Periodic Labs — Member of Technical Staff

  Sep 2025 – present
  • Develop high-throughput density functional theory workflows and machine-learning force fields for atomistic simulations.

• Radical AI

  Apr 2024 – Jun 2025
  • Worked on machine-learning interatomic potentials and robotic laboratories.

• Materials Project

  Jul 2022 – Dec 2023
  • Contributed to CHGNet and MACE-MP foundation models for atomistic simulations.
  • Developed high-throughput workflows generating reference datasets for interatomic-potential training.

  Selected Open Source All projects →

• Matbench Discovery

  251 531 commits Python, TypeScript, Svelte

  Maintainer. Developed a benchmark evaluating machine-learning models in realistic crystal-discovery campaigns.

• TorchSim

  488 39 commits Python

  Maintainer. Co-develop a PyTorch-native engine for batched atomistic simulations.

• CHGNet

  404 196 commits Python, C, Cython

  Maintainer. Helped develop a charge-informed interatomic potential.

• MACE Foundation Models

  302 1 commits Python

  Contributor. Helped train and evaluate MACE-MP for atomistic simulations.

• pymatgen

  1961 1050 commits Python, Cython

  Maintainer. Python materials-analysis library powering the Materials Project.

• atomate2

  347 386 commits Python, Jupyter Notebook, Shell

  Maintainer. Implemented MatPES PBE and r2SCAN workflows generating reference data for universal interatomic potentials.

  Selected Publications

1. Matbench Discovery - A framework to evaluate machine learning crystal stability predictions

   J. Riebesell, R. Goodall, ..., K. Persson  —  Nature Machine Intelligence  —  Jun 2025  · DOI  —  5 citations

2. TorchSim: An efficient atomistic simulation engine in PyTorch

   O. Cohen, J. Riebesell, ..., A. Gangan  —  IOP AI for Science  —  2025  · DOI  —  2 citations

3. CHGNet as a pretrained universal neural network potential for charge-informed atomistic modelling

   B. Deng, ..., J. Riebesell, ..., G. Ceder  —  Nature Machine Intelligence  —  Sept 2023  · DOI  —  397 citations

4. A foundation model for atomistic materials chemistry

   I. Batatia, ..., J. Riebesell, ..., G. Csányi  —  arXiv.org  —  Dec 2023  —  171 citations

5. Discovery of high-performance dielectric materials with machine-learning-guided search

   J. Riebesell, T. Surta, ..., A. Lee  —  Cell Reports Physical Science  —  Oct 2024  · DOI

  Education

• PhD Student - University of Cambridge

  Thesis: Towards Machine Learning Foundation Models for Materials Chemistry  

• MPhil in Scientific Computing - University of Cambridge

  Thesis: Probabilistic Data-Driven Discovery of Thermoelectric Materials

• MSc in Physics - ITP Heidelberg

  Thesis: Functional Renormalization Group Analytically Continued to Finite Temperatures

• BSc in Physics - Hamburg University

  Thesis: van der Waals Corrections for Density Functional Theory - DFT+D2 applied to Graphene-hBN-Heterostructures

  Technical Skills

Scientific computing

• Python
• Rust
• PyTorch
• JAX

Tooling

• Git
• Slurm
• Node.js
• Deno
• Vite
• PlayWright

Web

• TypeScript
• Svelte
• React
• REST
• MongoDB

  Community

• Associate Editor IOP AI for Science 2025 - present
• Nature Computational Science reviewer 2025 - present
• Digital Discovery reviewer 2024 - present
• NPJ Computational Materials reviewer 2024 - present
• Cambridge Physical Society 2020 - present
• Materials Research Society 2023 - present
• Materials Project Software Foundation 2023 - present

  Hobbies

• photography
• hiking
• cycling
• climbing