what happens at companies where there a LOT of math. Like fintech, AI, etc companies.
Are there equivalents of numpy,pandas, etc for go ?
what happens when you need to deploy models ?
Are there equivalents of numpy,pandas, etc for go ?
what happens when you need to deploy models ?
It turns out lots and lots of systems shuffle bytes & text so golang has a lot of room for adoption but for now modeling, business logic, math etc are all better done in other languages.
Maybe TensorFlow will become more used which would allow more cross language (and thus go could be added to this domain) but for now its a pretty bare stack.
Numpy equivalents: Gonum (gonum.org) and/or Gorgonia Tensors (https://gorgonia.org/tensor). Here's a cheatsheet for when to use what: https://www.cheatography.com/chewxy/cheat-sheets/data-scienc...
Pandas equivalents: gota (https://github.com/go-gota/gota) or qframe (https://github.com/tobgu/qframe).
I deploy deep learning models with Gorgonia (https://gorgonia.org/gorgonia). I have found them to usually be faster than TensorFlow or PyTorch. CUDA is a bit iffy, but works, kinda. Distributed learning, you gotta build your own which is fairly straightforwards in the library.
Compile them for your target. Move the binaries to your required server. Your CI/CD tool can do that easily
I still find myself reaching for R more often for stats, or Python more often for quick systemsx-ish things that need a little numerical work, but overall I give Julia high marks and hope it continues to gain traction.