I've been working through the Statistical Rethinking[0] course with Julia recently and so far the language has been really intuitive, same for the libraries (Turing, Distributions). And it's nice that you don't have to vectorize your code and can use normal for-loops.

Also, the way dynamic dispatch + specialization work is beautiful.

[0]:https://github.com/rmcelreath/stat_rethinking_2022