40 Worked Steps · Instant PDF Access · Arithmetic You Can Follow
The Network Room

Start Here

Everyone says it learns. Nobody shows you the arithmetic.

Forty steps from a single multiplication to a network that actually trains. Every step is small enough to work through on paper first, then in nine lines of code — the forward pass, the loss, the gradient, the update, and why each one is shaped the way it is.

$10.00One-time
A Network Is Just Arithmetic That LearnsInstant PDF · 40 worked entries
Yes, show me the arithmetic — Only $10.00

Secure checkout · Instant download · 60-day guarantee

See what's inside ↓

Forty steps, each one doable by hand before you write a line of code.
A Network Is Just Arithmetic That LearnsFrom the book
A 52-page PDF you work through, not a video you watchWhat you get
40 stepsEach with worked numbers
6 chaptersNeuron to training loop
Instant accessEmailed the moment you buy
Any devicePhone, tablet, laptop or print

Forty worked steps across six chapters — the neuron, the layer, the loss, the gradient, the loop and the first real network.

  • Every shape written out, at every layer
  • Code in plain Python and NumPy — no framework needed to follow it
  • A numeric gradient check you can run against your own derivation
  • Printable: the worked pages are meant to be marked up

Sound familiar?

You can run the tutorial. You still can't say what it did.

You copied the notebook, the accuracy went up, and nothing inside your head changed. The words are familiar — tensor, gradient, backprop, epoch — but if someone asked you to draw what happens to one number as it moves through one layer, you would stall.

That gap is not a lack of talent and it is not fixed by another framework tutorial. It is fixed by doing the arithmetic once, small enough to hold in your head, until the big version is obviously the same thing repeated.

Sample pages

A look inside

Every step is laid out the same way: what it computes, what you need, the worked numbers, and the mistake that makes it look like magic instead of arithmetic.

Step 03 · One neuron, by hand

Weights, Sum, Bias, Squash

  • Three inputs
  • Three weights
  • One bias
Time 20 minDo it on paper first

Step 14 · Where the learning is

The Gradient Of One Weight

  • The loss
  • One weight
  • A small nudge
Time 35 minCheck it numerically

Step 27 · The whole loop

Forward, Loss, Backward, Step

  • Nine lines of code
  • One toy dataset
  • A printed loss
Time 45 minWatch the loss fall

40 worked steps in total

Every step carries its numbers — the shapes, the values, the derivative — so nothing rests on a phrase you are supposed to already understand.

What's inside

Here's exactly what you'll be able to do

Why this works

It builds the whole machine from one multiplication, in an order where nothing appears before the thing it depends on.

Small Enough To Hold

Every example uses numbers you can carry in your head — three inputs, two layers — so the arithmetic stays visible instead of disappearing into a GPU.

Paper First, Code Second

You do each step by hand before you see the code. The code then reads as a transcription of something you already understand.

Nothing Skipped

No step says 'the framework handles this'. Where the framework handles it, the book shows what it is handling.

What you get

Everything in this book, listed

A 52-page PDF you work through, not a video you watchWhat you get
40 stepsEach with worked numbers
6 chaptersNeuron to training loop
Instant accessEmailed the moment you buy
Any devicePhone, tablet, laptop or print

Forty worked steps across six chapters — the neuron, the layer, the loss, the gradient, the loop and the first real network.

  • Every shape written out, at every layer
  • Code in plain Python and NumPy — no framework needed to follow it
  • A numeric gradient check you can run against your own derivation
  • Printable: the worked pages are meant to be marked up

The offer

Stop running other people's notebooks.

A Network Is Just Arithmetic That Learns builds the whole machine from one multiplication — the forward pass, the loss, the gradient and the update — in forty steps small enough to do by hand before you write any code.

Yes, show me the arithmetic — Only $10.00

Instant digital download. One-time payment of $10.00. Educational material about how neural networks work — not a course, not a certification, and not a promise about employment or salary.

The 60-day guaranteeRead it, run one of the worked examples, and if it does not make the inside of a network clearer than the tutorial you abandoned last week, email us within 60 days for a full refund. No questions, no hard feelings.

Questions

Before you buy

Do I need to be good at maths?

You need to be comfortable multiplying and adding, and willing to follow a derivative being worked out slowly. The book does not assume you remember calculus — where a derivative is needed, it is derived in front of you with numbers attached.

Which framework does it use?

None, for the first five chapters. Everything is plain Python and NumPy so that nothing is hidden. The last chapter shows the same network in PyTorch so you can see which parts the framework was doing for you.

I've already trained models. Too basic?

If you can already derive a gradient by hand and predict every shape in a model definition, yes. If you have trained models that worked without being able to say why, this is the layer underneath the one you have.

Is this a course with videos?

No. It is a PDF you work through with a pen. There are no videos, no cohort and no certificate — and nothing here is a qualification.

Will this get me a job in machine learning?

No book can promise that and you should distrust any that does. This explains how the arithmetic works. What that is worth to an employer depends on far more than one PDF.

Is this a physical book?

No — it's a digital guide (PDF), emailed to you right after purchase. Nothing is shipped.

What if it isn't for me?

You're protected by a 60-day money-back guarantee. Email us within 60 days for a full refund — no questions asked.

Do the arithmetic once

Forty steps from one multiplication to a network that trains — small enough to follow, complete enough to build on.

$10.00One-time
A Network Is Just Arithmetic That LearnsInstant PDF · 40 worked entries
Yes, show me the arithmetic — Only $10.00

Secure checkout · Instant download · 60-day guarantee

This is educational material about how neural networks work. It is not a course, a certification, a bootcamp or a career programme, and it makes no promise about employment, salary or professional outcomes of any kind. The code and the worked figures are teaching examples, written for clarity rather than for production: read them, adapt them, and test anything you reuse. Library APIs change often, so the method is what carries over, not the exact call signature. Nothing here is professional, legal or financial advice. Models reproduce the patterns and the biases of the data they are trained on; deploying one that affects people carries responsibilities this book does not cover.