|Abstract:|| Numerical meshes are essential components for forecasting the physical behavior of subsurface domains. Structured grids are often used for this purpose, but they lack flexibility to accurately represent complex subsurface geometry and to provide local level of detail in heterogeneous media. Unstructured meshes can, in principle, address these two gaps. In spite of much progress, challenges still exist to easily create unstructured meshes whose features are compatible with geological heterogeneities, with the physics to be simulated and with the selected numerical method. In this talk, I will review some recent and ongoing work RING to address these problems and discuss some of the remaining challenges. I will also highlight some features of the RINGMesh library, which is as a platform to address mesh-related problems.