Fellowship project title: Backend-agnostic parallel programs in Python
The goal: Make it easier for scientists to work on complicated climate models by using good software practices and hiding the most complicated part from them.
The why: Writing programs that are suitable for running efficiently on a supercomputer is often difficult, and fact that there is often a lack of continuity in scientific software development makes it even more complicated. We work on developing software that hides the complicated parallelization part from scientists that work on scientific models. This way they can fully focus on making their model work correctly on their own machine, e.g. a laptop. Moreover, good software practices can help with the continuity problem.