Latest

A narrative tour of a single token — Pi — from raw text through tokenization, special-token wrapping, the forward pass, scratchpad reasoning, and output stripping. With the Life of Pi framing.

Embeddings, projection, Q/K/V, dot product, softmax, weighted sum, positional encoding, matrix shapes, and the intuition behind attention — built up from first principles.

Learn the fundamentals of audio: how sound works, digital audio representation, sampling theory, and audio formats. Build a strong foundation for audio deep learning.

Master essential signal processing techniques for audio: Fourier transforms, spectrograms, mel-scale, MFCCs, and other feature extraction methods crucial for deep learning.

Bridge audio processing and deep learning: understand neural network architectures for audio, build your first audio classifier, and learn best practices for audio ML.