Since ~2003 or so. I’ve worked on various different types of projects, from small games using PyGame, via Django-based website backends (for CUYC), to my day-to-day work in scientific computing (NumPy, SciPy), visualisations (Matplotlib & co) and machine learning research and applications (TensorFlow, PyTorch, JAX, etc.). I’m one of the core developers of GPflow and GPflux.
On and off since ~2013. Lots of little toy projects and demos to investigate specific problems. I co-developed the original MLKernels.jl package and, more recently, have been contributing to various JuliaGaussianProcesses projects.
2009-2016. I first picked it up for my Master’s project. Modern Fortran is actually a very decent language! And has its deserved use-cases in scientific computing. One of the bigger projects I’ve contributed to are various parts of the libAtoms/QUIP code, including MPI for parallelisation on High Performance Computing clusters.