Agentic AI

Across five modules it builds the four core agent design patterns — Reflection, Tool Use, Planning, Multi-Agent Collaboration — from first principles in Python, before touching any framework. That sequencing is the whole value: understanding the patterns themselves means the framework-specific courses later make sense instead of feeling like magic incantations. It's a foundations course, not a production deep-dive.

So it's the ideal first step into agents, and not what you want if you already know the patterns and need deep, framework-specific production skills. It's free on DeepLearning.AI's own platform, certificate included, though you may spend a little on model API calls while building. A credible starting signal (Ng's name), but a foundation to build on rather than a job qualification by itself (as of 2026).

Agentic AI is Andrew Ng's course on building AI systems that take action through iterative, multi-step workflows. Across five modules it teaches the four agentic design patterns — Reflection, Tool Use, Planning and Multi-Agent Collaboration — building each from first principles in Python before layering on frameworks, so you understand what's happening under the hood.