You got the discussion going Naser ;-) Hope I am not being too annoying by being the Nth person to mention Julia here.

But I think that in this day and age, it would not be fair to talk about data science and optimization without mentioning Julia.

I have spent something like 15 years of my life writing C++ code, and it was not a great experience. It is a langauge which is only growing more complex every year. For data scientists who simply want to focus on solving their problem, I think it is asking a lot to torture them with a language which just has and endless list of ways of shooting yourself in the foot. Epically obscure error message. "Fun" undefined behavior combined with ming boggling complexity.

While you may be able to squeeze out most performance from C++, it is simply unreasonable to ask people to learn C++ to solve these kinds of problems today when Julia is available.

That is a language far more friendly to Python developers. It is just as easy to use (I would claim easier) and out of the box you get code running at least 10x faster than Python. With some optimization effort you can get very close to C++ code.

Developer time is a currency. You can pay your time doing different things. The time you spend setting up a C++ development environment, learn the language, troubleshoot problems etc, is time better spent optimizing Julia code.

I am still puzzled that a language which is such a game changer, is still so frequently ignored. If you want to make life easier for data scientists, then it is a no brainer solution.