Development of accurate numerical simulation methods for applications in computational materials science, as well as computational physics and chemistry. He is involved in the development of algorithms for First-Principles Molecular Dynamics, a simulation method that combines a calculation of atomic trajectories with a quantum mechanical description of electronic properties. This method is widely used to explore the properties of solids, liquids, nanoparticles and biomolecules. First-principles simulations are computationally expensive and require the use of large parallel computers. The integration of efficient parallel numerical algorithms in simulation codes is an important part of his research.