
- Home
- Books
- Artificial Intelligence & Machine Learning
- Deep Learning & Neural Networks
- Grokking Deep Learning
Grokking Deep Learning by Andrew Trask is a foundational guide to building neural networks from scratch using pure Python and NumPy arrays. Many engineers can call high-level frameworks without understanding the calculus and matrix mechanics running beneath the code. This book removes that abstraction by building, training, and troubleshooting neural networks line by line.
What you will learn
- Construct forward propagation passes and calculate loss on real-world inputs
- Implement gradient descent optimization to update weights across neural layers
- Derive and write backpropagation logic from scratch using pure NumPy arrays
- Mitigate overfitting in deep architectures using regularization and dropout
- Build convolutional neural networks for image classification without external libraries
- Implement recurrent neural layers for sequence modeling and language generation
Who this is for
Software engineers, students, and data practitioners who want to understand deep learning from first principles. Basic familiarity with high school mathematics and fundamental Python syntax is all that is required.


