This is a demo of how a machine learning unifying engine would work on the minimal level with some examples. Much like micrograd by Andrej Karpathy is inspired by Pytorch's Autograd, this is inspired by the repository unifyai/ivy created by Daniel Lenton.
The idea is to allow framework-agnostic code written once that can be run just by selecting a backend framework to run code with. One may check the examples for more clarification.
Please check Ivy if you require more comprehensive usage.