When I started learning Artificial Intelligence, I found hundreds of tutorials — but few structured paths that took you from "what is a neural network?" to "I can build and deploy a computer vision model." Artificial-Intelligence-A2Z is the curriculum I wish I had: a progressive, hands-on journey through AI, Machine Learning, and Computer Vision, built entirely in Jupyter Notebooks.
Why Build a Curriculum?
The AI learning landscape in 2024–2026 is overwhelming. Between LLM hype, endless YouTube courses, and academic papers, it is easy to jump into advanced topics without solid foundations. I created A2Z to address three gaps:
- No prerequisites assumed. Basic Python is helpful, but the notebooks explain mathematical concepts as they arise.
- Hands-on from day one. Every concept is accompanied by runnable code — not just theory slides.
- Progressive complexity. Each notebook builds on the previous one, creating a coherent learning arc.
Curriculum Structure
The repository is organized into three major tracks:
Track 1: Foundations of AI & ML
- Introduction to AI, ML, and Deep Learning concepts
- Linear algebra and statistics refresher (only what you need)
- First models: linear regression, logistic regression, decision trees
- Model evaluation: accuracy, precision, recall, F1, cross-validation
Track 2: Deep Learning
- Neural network fundamentals with NumPy (before touching frameworks)
- PyTorch basics: tensors, autograd, training loops
- Convolutional Neural Networks (CNNs) for image classification
- Transfer learning with pre-trained models
Track 3: Computer Vision
- Image processing with OpenCV
- Object detection concepts and YOLO introduction
- Face recognition pipelines
- Real-world CV project: building an end-to-end detection system
Design Principles
Notebook-First Learning
Jupyter Notebooks are the ideal medium for AI education. Learners can read explanations, run code cells, modify parameters, and immediately see results. Each notebook follows a consistent structure: concept introduction, mathematical intuition (where needed), code implementation, exercises, and further reading.
Build Before Abstract
Instead of starting with backpropagation equations, the early notebooks have learners train a model and observe results first. The math comes after the intuition — mirroring how most engineers actually learn.
Open and Iterative
The repository is public on GitHub and evolves over time. As I learn new techniques or find better explanations, notebooks get updated. Community contributions and issue reports help identify confusing sections.
Who Is This For?
- Developers transitioning into AI/ML roles
- Students looking for a structured supplement to coursework
- Self-learners overwhelmed by the fragmented AI tutorial ecosystem
- Anyone who learns best by doing, not just reading
Lessons from Building the Curriculum
- Scope control is hard. AI is vast — resisting the urge to cover everything and instead building depth in ML and CV was the right call.
- Runnable > Comprehensive. A notebook that runs end-to-end with synthetic data is more valuable than one referencing datasets the learner cannot access.
- Teaching reinforces learning. Writing these notebooks deepened my own understanding of concepts I thought I already knew.
Get Started
Clone the repository, open the first notebook, and start learning. No setup beyond Python, Jupyter, and the listed dependencies.
Explore the full curriculum on GitHub.