Talk as part of a colloquium in Data-Driven Materials discovery.

Summary: Materials informatics has become an important discipline across chemistry and engineering that offers efficient and effective pathways for discovering and targeting the design of materials tailored to specific applications. While traditional computational methods often scale poorly when navigating the vast configurational space of complex systems, data-driven approaches provide an efficient, scalable alternative. Here, I present how machine learning architectures and chemical foundation models are applied directly to predictive property mapping, inverse design, and screening candidates for sustainable technologies and energy storage. The presentation details computational pipelines that accelerate the materials discovery, demonstrating how informatics bridges raw data with actionable materials innovation. Ultimately, the modular nature of these data-driven frameworks highlights their adaptability across diverse material classes, establishing a versatile paradigm for next-generation discovery.