This project applies data-centric AI to address a key challenge in chemical research: the scarcity of structured, machine-learning-ready data. A significant bottleneck in the field is that vast amounts of chemical information—spanning small molecules, proteins, reactions, and knowledge graphs—remain locked in unstructured formats within literature and databases.

The SUDS Scholar will learn and apply state-of-the-art data extraction and curation techniques to create foundational datasets. A core focus will be pioneering modern data documentation practices not yet widely adopted in chemistry. The student will create comprehensive data cards and model cards to ensure transparency and responsible use. Furthermore, they will format these datasets using the Croissant ML metadata framework, a new standard for making datasets discoverable and ‘ML-ready.’This work will directly contribute to accelerating AI-driven discovery in chemistry and provide the student with unique, hands-on experience at the intersection of data science and chemical research.