Also you don't need calculus to "do" ML (even deep learning research!) I got to the point of writing my own toy neural network from scratch, seeing backpropega

Also you don't need calculus to "do" ML (even deep learning research!) I got to the point of writing my own toy neural network from scratch, seeing backpropegation, figuring that I'd have to use the chain rule myself on my forward pass, understanding what "automatic differentiation" was and why it's important, and decided "screw that I'm not putting myself through this hell again" and decided to look into https://en.wikipedia.org/wiki/Derivative-free_optimization https://en.wikipedia.org/wiki/Derivative-free_optimization You can literally just find your own hetrodox subset of scholars in your field who are like "calculus? pfft!"