Gradient Descent

Gradient Descent

From Hand Calculations to AI-Assisted Code

You can already ask an AI to write you a Gradient Descent routine. The problem is knowing why it's diverging to infinity, crawling so slowly it looks frozen, or settling confidently into a local minimum that isn't the answer — all three usually trace back to a single number most people set by guesswork. This is the course that gives you that judgment.

Five parts, one algorithm, no gaps:

• Part 1 — Intro. How following the slope downhill, one step at a time, minimizes a function — and why this one idea sits underneath nearly everything in modern machine learning. The intuition before the notation.

• Part 2 — By Hand. A complete worked example, step by step: compute the gradient, take a step, watch the value drop — every partial derivative and every number shown. Nothing waved away, nothing left "as an exercise." You prove the algorithm to yourself.

• Part 3 — In Excel. The same example built out in a spreadsheet you can open and poke at, so you watch it converge — and see for yourself exactly what a too-large or too-small learning rate does to the path.

• Part 4 — In Python. A clean, executable, verified reference implementation — the known-good version you can check any other code against.

• Part 5 — With AI. A tested prompt that directs an AI assistant to carry this same algorithm to your problem, plus the understanding to catch it when it's wrong.

By the end you won't just have working code. You'll have followed the algorithm from first principles through a hand calculation, a spreadsheet, and production Python — so when you point an AI at your own optimization problem, you'll know how to set the learning rate on purpose, recognize divergence and local minima on sight, and tell when the answer is right.

Written for practicing engineers and STEM professionals who'd rather understand a method than trust a black box.