Hi

I’m research scientist in Meteorology. Since December 2025, I’m SMASH fellow at the University of Ljubljana, developing data assimilation (DA) methods for coupled Earth system models. In order to make predictions on medium to long timescales, classical approaches use semi-empirical transformations with analytical error relationships. However, classical approaches struggle to capture nonlinear and multiscale error structures, limiting analysis and forecast accuracy. We address these limitations by introducing a neural network-based covariance model with a data-derived error structure. My research leverages variational autoencoders (VAEs) to transform innovations, prior errors, and observation errors into a smaller-dimensional, nearly Gaussian latent space, which improves assimilation performance. Thus, this project is in a position to advance weather and earth-system forecasting, benefiting sectors like agriculture, and energy.

In my PhD (2020-2024), I studied the assimilation of cloud-affected satellite visible and infrared observations to improve convective-scale numerical weather prediction (i.e. small-scale storms).

Selected talks:

• 06/2026: MLDADS workshop (Hamburg, Germany)
• 06/2024: Ambiguity and nonlinearity in the assimilation of visible and infrared observations. ISDA seminar series (online), see the recording

See my latest publications. See my CV, PhD thesis, MSc thesis, and BSc thesis.

Disclaimer

Co-funded by the European Union. Views and opinions expressed are however those of the author(s) only and do not necessarily reflect those of the European Union or European Research Executive Agency. Neither the European Union nor the granting authority can be held responsible for them.

Last edit: 15 July 2026