Efficiently harvesting light as a direct means of energy production and as an aid to catalyze chemical reactions is a significant technological challenge riddled with many complexities. There exists a critical computational need for software that can simulate these light-matter interactions and fundamentally understand charge- and energy-transfer processes in photovoltaics, light-sensing devices, photocatalysts, and light-activated proteins, with an ultimate goal of optimizing materials for diverse optoelectronic applications or understanding biomolecular processes within the life sciences. Light-matter interactions are probabilistic events and a significant computational challenge is that massive amounts of data must be generated to explore these fundamental processes. The primary objective of this computational chemistry sciences team is to design a machine-learning, molecular-dynamics environment that will utilize current petascale and future exascale computational capabilities to advance understanding of charge and energy flow in materials. Simulation tools to evaluate charge- and energy-transfer pathways have significantly progressed as efforts have been made to improve computational methodologies. The developed machine-learning environment will 1) comprehensively advance both a technical and conceptual understanding of light-matter interactions as applied to a variety of systems including photovoltaics, light-sensing devices, photocatalysts, crystal-dye interactions, and light-activated proteins; 2) produce transformative algorithms and the developed machine-learning methodologies are knowledge gained by the scientific community as we release software and algorithms to the scientific community for general use; 3) have direct impact in advancing the rapid discovery of new photo-activated materials as fast and efficient algorithms are needed to increase the pace of materials discovery; and 4) boost interdisciplinary collaborations within the materials and chemical sciences communities.