I am a PhD student, working on Sensor Fusion and estimation problems. I would like as I finish my PhD to have acquired sound knowledge in what seems to be the industry norm for working with real perception systems (irregardless of context) , modern C++. This is a required skill in literally any job post that I see. I do have basic knowledge and experience of how it works, but it just feels simpler, more time efficient, easier to visualise, and easier to debug for a researcher to test an algorithm on Python or Matlab. I totally understand the benefits of using a scripting language, but I feel that we as engineers are over-relying on such tools.
Say for instance you want to test a RANSAC algorithm. My go-to approach would be start writing lines in MATLAB, no compilation necessary, all the tools are available in a single-line call, immediate access to documentation, very informative error messages, customisable plots on-the-fly and the list grows on. At the same time, I don't even want to imagine how much time it would take me to achieve the same results in C++, while most of the time I would have to spent it fighting with segmentation errors. So this brings me back to my question: How do I, in the context of my everyday PhD life, systematically build sound software development experience in C++, within my domain of expertise?