Since the blog post does not have many code samples, this non-trivial AD example with Zygote.jl is worth sharing (it's from their readme):

    julia> using Zygote
    julia> fs = Dict("sin" => sin, "cos" => cos);
    julia> derivative(x -> fs[readline()](x), 1.0)
    cos
    -0.8414709848078965
    julia> -sin(1.0) # 'true' derivative in 1.0
    -0.8414709848078965

So Zygote can apply AD to an anonymous function that looks up a function in a hash table from user input.

https://github.com/FluxML/Zygote.jl