I don't know what OPs project looks like, but in our company the choice has a clear path: - performance/low level: Rust, Cpp
- CPU/memory is not the bottleneck and I don't need much concurrency / it's data science, machine learning, ETL ... : Python
- concurrency is key, highly distributed system: elixir, golang, erlang
- programming based on mathematical models: Haskell
Check this out: https://lispcookbook.github.io/cl-cookbook/numbers.html
Can you elaborate the cons that made you decide to move away from Python, C++, and Julia?
Julia makes sense to me, as the youngest production language and community.
But Python? It is the #1 most popular programming language globally. C++ is also very popular. This means you can easily find people to hire globally (making business easier and more economical). You can also easily find code doing almost anything you want---speeding up time to value and time to ship.
I'm wondering what cons are driving your decisions to move on from Python and C++?
A short summary of my feelings is that it has some syntactic oddities, it's too incomplete in some areas to fully evaluate (particularly error handling), and it completely botches overflow. Whatever your opinions on undefined behavior on integer overflow, there really is no justification for keying undefined-versus-well-defined on signed versus unsigned.
Given that Carbon isn't ready, I think if you're porting something from C++ and Carbon looks good to you, Rust might be the way to go.
Alternatively, if you wanted something more functional, I've heard good things about OCaml