Markus Diesmann is interested in models of the brain at the resolution of nerve cells. The construction of brain-scale models requires the integration of anatomical and physiological data. This reveals consistency issues and uncovers gaps in knowledge. However, model studies are often not reproducible, challenging a fundamental principle of science: the ability to build upon the work of others. Diesmann advances the digitalization of workflows and concepts for the reusable specification of models. The conceptual separation of models from generic simulation codes is an achievement of the community. Different models, also future ones, can be simulated with the same simulation engine. This enables the maintenance of reliable and optimized codes. Diesmann is a founder of the NEST code and contributes to the establishment of simulation as a third pillar of science next to experiment and theory. The NEST Initiative follows an open source and community strategy. NEST scales from laptops to exascale supercomputers and is a reference for the verification and validation of neuromorphic systems. A model maintained by Diesmann and colleagues, covering a square millimetre of brain surface has become a de facto standard benchmark. In 2001 Diesmann and colleagues realized that software development in science is necessarily different than in industry and have been working on suitable methods and software architectures since. Diesmann and colleagues are developing strategies for research organizations for both Research Software Engineering (RSE) and the operation of scientific software as infrastructure. Diesmann promotes an infrastructure-driven approach to science as a complement to the hypothesis-driven approach dominating neuroscience. He criticizes the fascination of computational neuroscience with minimal models, as it remains unclear how they can be combined with each other, or iteratively integrate new data to achieve a larger explanatory scope. Diesmann is interested in science strategy and policy as well as advanced education. He coordinated the development of neuromorphic computing as a new strategic field at Research Centre Juelich, and guided a pilot project for its build-up. Diesmann heads JuDocS, the central organization for 1200 doctoral researchers and 400 supervisors at Research Centre Juelich. His institute is committed to equal opportunities, a family friendly environment, and collegial leadership. Three of the four departments of the institute of Diesmann are led by female scientists, and Diesmann co-leads the institute together with Sonja Gruen.