Olga Mironova
Olga Mironova holds degrees in Mathematics, Economics, and Data Science (Paris Lodron University of Salzburg). Prior to joining research, she gained industry experience applying machine learning to sales forecasting and advertising evaluation at a leading FMCG company. She currently works as a Research Assistant with the Smart Analytics and Reinforcement Learning team, contributing to the INSPIRE project on RL-based personalised diagnostics in STEM primary education. Her broader research interest lies in bridging classical control and reinforcement learning through structured priors for robust control of constrained physical systems (application to beam steering at the CERN AWAKE experiment).
Olga Mironova

Position: Research assistant: (INSPIRE)
E-Mail: olga.mironova@plus.ac.at
Datum Beginn: 15.02.2024
Datum Ende: 28.02.2027
Publikationen
- S. Hirländer, O. Mironova, S. Trausner, L. Grech, L. Fischl, A. Santamaria Garcia: Koopman-Stabilised World Models for Offline Reinforcement Learning in Accelerator Control. (2026) https://doi.org/ https://doi.org/10.18429/JACoW-IPAC2026-WEP6098
- S. Hirländer, K. Björkbom, S. Trausner, O. Mironova, L. Fischl, P. Auer, R. Ortner, V. Kain: Reinforcement Learning Beyond Greedy Optimisation for Accelerator Control with Delayed Consequences. (2026) https://doi.org/https://doi.org/10.18429/JACoW-IPAC2026-WEP6097
- S. Hirländer, O. Mironova, S. Trausner, L. Fischl, T. Gallien, L. Grech: Causal GP-MPC: Where Structure, Safety and Online Learning Come Together for Robust Accelerator Control. (2026) https://doi.org/https://doi.org/10.18429/JACoW-IPAC2026-WEP6096
