Why I Built a 3,200-Line Python Pipeline to Generate Synthetic Financial Data From Math -- Not AI

Most synthetic data tools feed real data into a neural network. I did the opposite: Pareto distributions first, algebraic constraints second, AI last. Here's why.