The calculus that trains every model — derivatives, gradients, and the optimization machinery behind gradient descent, backpropagation, and loss minimization.
