Classifies AI two ways — by capability (Narrow, General, Superintelligence) and by function (reactive through agentic) — and explains why today’s most advanced systems are still advanced narrow AI wrapped in agentic runtimes, not AGI.

Explains how deep learning sits inside machine learning, compares the two on data, compute, interpretability, and use case, and shows how foundation models, TinyML, and hybrid designs blur the old dividing lines.

Breaks the modern AI system into five layers — infrastructure, data and development, foundation models, serving and orchestration, applications and agents — and explains the two shifts reshaping it: inference economics and the agentic runtime.

Introduces NLP, the AI discipline that lets machines read, interpret, and generate human language. Covers the processing pipeline, the NLU/NLG split, common applications, the arc from rules to transformers, and NLP’s role as the intent engine for agentic AI.

Explains embeddings — numerical vectors that capture meaning — and the vector databases built to store and search them, and shows why together they power semantic search, Retrieval-Augmented Generation, and long-term memory for AI agents.