Hi, I'm

Tomasz Różański.

I am a Postdoctoral Researcher at the Australian National University, where I work on machine learning for astrophysics, with a focus on stellar spectroscopy. I completed my PhD in astrophysics at the University of Wrocław in Poland, building on a background in computer science, astronomy, and mechatronics.

My research develops methods for measuring stellar properties from large spectroscopic surveys, with the aim of making these measurements more accurate, precise, automated, and scalable. I work on problems ranging from data reduction and radiative-transfer modeling to interpretable machine-learning emulators and statistical inference methods. More broadly, I am interested in machine learning for science and in building scientific software that connects simulations, models, and observations.

Feel free to reach out if you would like to discuss research, software, or collaborations.

email: tomasz.rozanski1@anu.edu.au

Portrait of Tomasz Różański

Selected papers

  1. Scaling Laws for Emulation of Stellar Spectra Open Journal of Astrophysics, 2025. arXiv abstract doi ads
  2. TransformerPayne: improving forward modelling of stellar spectra with transformers The Astrophysical Journal, 980, 66, 2025. arXiv abstract ads
  3. SUPPNet: neural network for stellar spectrum normalization Astronomy & Astrophysics, 659, A199, 2022. arXiv abstract doi ascl github ads

A complete list of publications is available on NASA ADS.

Apps

Scientific and educational: