Assumed background
No advanced math is assumed. You should be comfortable with basic arithmetic and willing to install and use Python; the book explains the data, metric and model ideas as they appear.
PDF book · AI engineering
AI Systems Craft
From NumPy and machine learning to neural networks, transformers and LLM apps, project by project
A guided AI engineering course in Python. Every idea is a runnable project with its exact output, from NumPy arrays and data preparation to neural networks, transformers and LLM applications.
- 854 pages
- PDF download
- 2026 Edition
- English
- $25 one-time
Secure checkout with Stripe. Your receipt gives you a private download link, and there is no subscription.
About the book
What AI Systems Craft covers
AI Systems Craft is a practical bridge from ordinary programming curiosity to building AI systems. It starts with NumPy arrays, broadcasting and linear algebra, pandas, and data preparation and feature engineering, then builds evaluation metrics, classical machine learning with scikit-learn, regression and classification from scratch, optimization and gradient descent, neural networks from scratch and PyTorch. The advanced chapters cover images as arrays, natural-language processing, embeddings and similarity search, transformers and attention, LLM engineering with prompts and retrieval-augmented generation, and scaling, compute, cost and MLOps.
Best for: Analysts, marketers and developers who want to understand and build with machine learning and LLMs through runnable projects rather than abstract jargon.
- 854 pages in 16 chapters, from NumPy fundamentals to scaling, compute, cost and MLOps
- 209 runnable Python projects, each with its code and exact output
- Classical machine learning, neural networks, PyTorch, NLP, embeddings, transformers and RAG
- Skill maps, worked traces, common mistakes, quizzes and answer keys for self-study
How it teaches
A repeatable study rhythm
Every project follows the same rhythm: read the data shape, name the transformation, predict the metric or output, then change one detail safely. Chapters open with a skill map and the idea behind the concept, work through short runnable projects with their exact output, worked traces, visual models and common mistakes, and close with a quiz and its answer key. Basic arithmetic and a willingness to follow a small program line by line are enough to begin.
Skill map
What the chapter adds, what it builds on and where it leads.
The idea behind it
The problem the concept solves, before any code.
Project
A complete runnable program with the exact output to check against.
Trace and model
Follow the setup, the transformation and the evidence that proves it worked.
Common mistakes
The traps learners fall into, and small changes to try next.
Quiz
A few questions at the end of each chapter, with full answers in its answer key.
Look inside
Pages from the PDF
Five full pages from the PDF. Select a page to open it at full size.
How to use this bookThe signposts every chapter repeats: the idea behind it, a project with its exact output, common mistakes, try-it changes and a quiz. 
ContentsSixteen chapters in three parts, from NumPy fundamentals to scaling, compute, cost and MLOps. 
A chapter openingChapter 1 begins with a skill map, the idea behind the code and a first program with its output. 
Inside a projectWhat you will build, a logic lens on each line and a visual model of setup, rule and evidence. 
Attention by handChapter 14 computes attention weights with NumPy, so a transformer's core step shows in the output.
Before you start
What you need and how long it takes
A realistic pace
A realistic self-study pace is 8 to 12 weeks if you work through several lessons a week and run the projects. Readers who already use Python can move faster through the early data chapters; beginners should treat the book as a lab manual: read, run, inspect and repeat.
- NumPy
- pandas
- scikit-learn
- Gradient descent
- Neural networks
- PyTorch
- NLP
- Embeddings
- Transformers
- RAG
- MLOps
One-time purchase
Get AI Systems Craft
$25USD · PDF · 854 pages
Buy the PDF · $25- Pay securely with Stripe. IndepApps never sees your card number.
- Your receipt page gives you a private, single-use download link.
- We email you a link to your receipt so you can find it again.
See the purchase terms and privacy policy. Questions? support@indepapps.com
- Format
- Pages
- 854
- Edition
- 2026 Edition
- Language
- English
- File size
- 16.4 MB
- Price
- $25 USD, one-time
- Delivery
- Single-use download from your receipt
Questions about AI Systems Craft
Is AI Systems Craft suitable for beginners?
No advanced math is assumed. You should be comfortable with basic arithmetic and willing to install and use Python; the book explains the data, metric and model ideas as they appear. It is written for self-study and builds up step by step.
How long does it take to work through AI Systems Craft?
A realistic self-study pace is 8 to 12 weeks if you work through several lessons a week and run the projects. Readers who already use Python can move faster through the early data chapters; beginners should treat the book as a lab manual: read, run, inspect and repeat.
What do I get, and how is it delivered?
An 854-page PDF (16.4 MB). After payment, your receipt page gives you a private, single-use download link, and we email you a link to that receipt. Save the PDF after downloading.
Can I read it on any device, and is it mine to keep?
It is a standard PDF for a computer, tablet or phone, online or offline. It is a one-time purchase of $25 with no subscription. You keep the file for your own reading; the purchase terms explain the personal-use licence.
What if my download fails?
If a transfer is interrupted, the link is released so you can try again from your receipt. If it still fails, email support@indepapps.com with your order number instead of buying again.
More PDF books
All books
Python Craft
A complete Python course built around runnable examples, exact output, construction habits, quizzes, and answer keys.
$25 · PDF
Java Craft
A complete Java course built around executed projects, exact output, traceable reasoning, and practical programming habits.
$25 · PDF
The Code Projects Companion
A large source-backed programming resource edition with project cards, logic exercises, code/output layouts, and cited source paths.
$25 · PDF