I am researching parallel algorithms for adaptive mesh refinement. Meshes are the basis of many common mathematical approaches to numerical simulation, helping to organize discrete function spaces, to visualize, load and store data, and to locate geometric objects. Mesh adaptivity refers to local refinement and coarsening, which accounts for spatial and temporal heterogeneity of the equations to solve. Adaptivity, simultaneously, introduces severe challenges to efficiency. My work is oriented towards developing algorithmic and software solutions that are fast in absolute terms and scale to the largest supercomputers worldwide. I have long been collaborating with geoscientists to support real-world simulations.