Research Topic:

Data-Driven Optimization Strategies for Materials Design

Supervisors:

Prof. Britta Nestler (KIT), Arnd Hendrik Koeppe PhD (KIT)

Thesis Abstract:

My PhD project develops an intelligent framework that helps material scientists steer their studies and reach reliable conclusions with fewer (costly) simulations and experiments. Bridging mathematics and computational materials science, the framework integrates Bayesian active learning, FAIR-aware research data management, interoperability across heterogeneous RDM systems, and scientific visualization to efficiently explore vast search spaces. Proof-of-concept applications from simulations and laboratory battery research demonstrate how the framework accelerates material design by supporting informed, uncertainty-aware decision-making.

Publications:

•Giovanna Tosato et al. “Accelerating battery research with an AI interfacebetween FINALES and Kadi4Mat”. In: arXiv preprint arXiv:2605.00909 (2026).

Conference Presentations:

•NHR Conference 2024 (Contributed talk + Poster)

• NHR4CES Community Workshop 2025 (Contributed talk)

•NHR Conference 2025 (Poster)