The PRECISE project (Physics-informed and data-driven Reliable Estimations for Corrosion In Sustainable nuclear Environments) advances nuclear materials degradation prediction, moving from empirical correlations to physics-informed and time-series-data based AI approaches, by developing a dual architecture that addresses both data-scarse and data-rich scenarios. The experimental program, coupled with features engineering, characterization and validation against both analytical and numerical methods, provides confidence in model predictions. The inclusion of data for different environments, cross-laboratory and cross-material validation further enhances the robustness and applicability of the developed methodologies.