HIIT Postdoctoral Fellow 1.2.2026-31.1.2029
Sara’s research focuses on Bayesian methods for learning and decision-making from sequential data. She develops scalable online inference methods to update beliefs about a system as new observations arrive, and studies how Bayesian experimental design can use those beliefs to choose informative experiments. Her current work focuses on dynamical systems, with particular interest in continual learning and how inference and experimental design can adapt when the underlying system changes over time.
Sara is an HIIT Postdoctoral Fellow in the Department of Computer Science at the University of Helsinki, where she is part of the Multi-source Probabilistic Inference group led by Prof. Arto Klami. She obtained her PhD from Universidad Carlos III de Madrid in 2022, working on statistical signal processing. Before joining Helsinki, she held postdoctoral positions at Aalto University and IMT Nord Europe, and was a visiting researcher at the University of Edinburgh.
Selected publications
[1] Pérez-Vieites, S., Iqbal, S., Särkkä, S., & Baumann, D. (2026). Online Bayesian experimental design for partially observed dynamical systems. International Conference on Machine Learning (ICML).
[2] Pérez-Vieites, S., Molina-Bulla, H., & Míguez, J. (2026). Nested smoothing algorithms for inference and tracking of heterogeneous multi-scale state-space systems. Foundations of Data Science, 8, 1–29.
[3] Cox, B., Pérez-Vieites, S., Zilberstein, N., Sevilla, M., Segarra, S., & Elvira, V. (2024). End-to-end learning of Gaussian mixture proposals using differentiable particle filters and neural networks. IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), 9701–9705.