Agentic AI

What the course covers

Agentic AI is about systems that do more than answer a single prompt. This course, taught by Andrew Ng, focuses on building agents that carry out work in repeated, multi-step cycles, which is a different engineering problem from one-shot generation.

Who it suits

The material is aimed at people who want to move from using language models to building software that plans and carries out tasks. Developers, product builders and technically minded managers are the natural audience, since the central question is how a system loops, checks its own work and decides its next step.

Key takeaways

• An agent is defined by repeated steps rather than a single call.
• Design choices made at each step shape how reliable the whole workflow becomes.

Expect a practical, build-oriented view rather than a survey of buzzwords.