A modern programming language that makes developers happier.

kotlinlang

i think it makes more sense to bind to the computation libs on the platform than rewrite everything in kotlin, so i'll look at how hard it is to get something like libdynd working with kotlin-native. unfortunately i dont think theres a wealth of C-only algorithms (gnu scientific library being the obvious candidate)

"wealth of C-only algorithms" - could you some examples of such ML or DA libraries?

My point is that all ML and DA projects I've seen were prototyped in Python/Mytlab/R, then rewritten in Java for production. From that I conclude that all core ML and DA libraries are duplicated on JVM

theres libraries like mlpack in C++ and deeplearning4j on the jvm. I'm not targeting ML yet (trying to implement a core numpy-like API as a first step) but if/when someone gets to implementing ML i think the right approach for a cross-platform ML lib in kotlin would be to bind to those libs and then you could write your stuff in kotlin and target them all

i agree the story is sparse on the jvm at the moment, but probably the best way to fix that is to contribute to the existing libraries like deeplearning4j