Abstract: Coarse-grained modeling is an essential technique for extending the time and length scales of molecular simulation and design. For molecular dynamic simulations, learning-based force fields have made significant progress in accelerating ab-initio MD simulation but are not fast enough for many real-world applications due to slow inference for large systems and small time steps (femtosecond-level). In this talk, we introduce a multi-scale graph neural network that simulates coarse-grained MD with a very large time step. This method demonstrates effectiveness in two complex systems: single-chain coarse-grained polymers and multi-component Li-ion polymer electrolytes, recovering structural and dynamical properties at several orders of magnitude higher speed than classical force fields. For materials design, Metal-organic frameworks (MOFs) are of immense interest in applications such as gas storage and carbon capture due to their exceptional porosity and tunable chemistry. However, designing MOFs poses challenges due to their intricate unit cell structures derived from molecular building blocks. We illustrate a novel approach that combines diffusion-based generative modeling and a coarse-grained representation based on the building blocks to design 3D MOF structures that are tailored for carbon capture applications.