I let experts write the high-performance code which uses a lot of computing resources (and hence cost/energy). However, there is a very significant level of software which exists "on top" of the quantum chemistry codes; automation tools and analysis scripts which manage high-throughput studies and extract useful information from the results. Of particular interest are properties that can be directly compared with experiment, for validation of theories and for interpretation of measurements.
I grew up in London and studied for almost ten years in the beautiful city of Bath, beginning with an undergraduate degree in Chemical Engineering and moving to the Centre for Sustainable Chemical Technologies for a PhD program. There was something addictive about the thrill of submitting a large parallel calculation to a remote cluster that drew me from a short Masters project into a full PhD project in computational materials chemistry and along the way a pile of Python scripts started to appear... Collaborative coding projects emerged naturally from research activities and provide a great incentive to work in a maintainable, considerate style. The first fellowship goal is to promote and assist good practice in materials chemistry groups, which largely develop their software in isolation. The second fellowship goal is to learn more about accessibility issues and develop guidelines and good practices for research software engineers; this is an area that is widely considered in the commercial and web domains but is potentially harmfully overlooked in the research world.