The project investigates theoretical and computational modeling of excited states and optical properties in molecular materials through hybrid methodological approaches. The scheme require first to provide a structures of molecular crystals or aggergate, either obtained from experiment or guessed via computational tools. Exciton model hamiltonians will be fed by excited-state calculations and ab-initio molecular dynamics, to derive hamiltonian parameters. The interface of ad-hoc developed machine learning (ML) algorithms will be also developed, allowing ML to improve the parametrization of the model hamiltonian, hence providing a full insight into excitons physics in an automatized fashion. The model will embed microscopical information that will translate into optical and exciton dynamical properties, to be finally compared and validated against experiments. Systems investigated will include innovative organic and bio- molecular materials, offering this project the chance to understand and control their diverse and promising optical properties.