Tulga-Erdene Sodjargal

Hello and welcome! I’m Tulga, a dude from Erdenet, Mongolia, working at the intersection of machine learning and computational chemistry. I’m currently working with Prof. Michele Ceriotti at EPFL, where I contribute to various projects related to integrating previously ignored long-range interactions into atomistic ML models. Also, I am working with Prof. Taras V. Pogorelov at UIUC, where I analyze the effect of sterol composition on lipid bilayer systems by translating statistical analysis results into biophysical insights.

My research interests include:

Long-range Interactions: Physics-based (inspired by current MD engines) and data-driven (ML-based) methods for adding long-range corrections to ML potentials
Statistical Methods for Biophysics: Advanced statistical techniques for analyzing molecular simulations
Scientific Software Development: Efficient, user-friendly, and robust software for computational chemistry

If you are interested in what I did and am doing, please don’t hesitate to contact me through whatever means listed on this website that are convenient to you.

news

Jul 25, 2025	Our preprint on design of a novel long-range machine-learned interatomic potentials, LOREM, is out on arXiv!
Jul 08, 2025	Our preprint describing a Python package for development of educational Jupyter Notebooks, scicode-widgets, is out on arXiv!
Mar 17, 2025	I presented the preliminary results of our trials of building long-range machine-learned interatomic potentials at DPG-Frühjahrstagung (DPG Spring Meeting) 2025 in Regensburg, Germany!
Sep 02, 2024	I am now an visiting undergraduate researcher at the COSMO lab at EPFL, Switzerland, working with Professor Michele Ceriotti on integrating long-range interactions into machine-learned interatomic potentials.

publications

arXiv

Learning Long-Range Representations with Equivariant Messages

Egor Rumiantsev, Marcel F. Langer, Tulga-Erdene Sodjargal, Michele Ceriotti, and Philip Loche

arXiv preprint arXiv:2507.19382, 2025

Bib HTML PDF

@article{rumiantsev2025learning,
  title = {Learning Long-Range Representations with Equivariant Messages},
  author = {Rumiantsev, Egor and Langer, Marcel F. and Sodjargal, Tulga-Erdene and Ceriotti, Michele and Loche, Philip},
  journal = {arXiv preprint arXiv:2507.19382},
  year = {2025},
  eprint = {2507.19382},
  archiveprefix = {arXiv},
  primaryclass = {chem-ph},
}

arXiv

scicode-widgets: Bringing Computational Experiments to the Classroom with Jupyter Widgets

Alexander Goscinski, Taylor James Baird, Dou Du, João Prado, Divya Suman, Tulga-Erdene Sodjargal, Sara Bonella, Giovanni Pizzi, and Michele Ceriotti

arXiv preprint arXiv:2507.05734, 2025

Bib HTML PDF

@article{goscinski2025scicodewidgets,
  title = {scicode-widgets: Bringing Computational Experiments to the Classroom with Jupyter Widgets},
  author = {Goscinski, Alexander and Baird, Taylor James and Du, Dou and Prado, João and Suman, Divya and Sodjargal, Tulga-Erdene and Bonella, Sara and Pizzi, Giovanni and Ceriotti, Michele},
  journal = {arXiv preprint arXiv:2507.05734},
  year = {2025},
  eprint = {2507.05734},
  archiveprefix = {arXiv},
  primaryclass = {physics.ed-ph},
}