Agentic AI Applications with Codex

About This Course

This course equips learners with practical skills to design, build, test, and optimise agentic AI applications using Codex. Participants will use Codex as an AI coding agent to translate application requirements into working software, understand existing codebases, generate and refactor code, diagnose issues, and validate solutions through testing and review.

Learners will explore the core components of agentic applications, including structured workflows, tool use, memory, context management, retrieval-augmented generation (RAG), action loops, human approvals, and multi-agent coordination. They will connect applications to external systems and data sources using Model Context Protocol (MCP) tools and develop reusable Skills for consistent, task-specific workflows.

The course covers the end-to-end development process, from defining use cases and planning system architecture to implementing interfaces, integrating APIs, managing data, and deploying functional applications. Participants will also use Codex to create tests, evaluate outputs, troubleshoot failures, improve reliability, and document their solutions.

Emphasis is placed on secure coding, permission control, data privacy, responsible AI practices, and human oversight. By the end of the course, learners will be able to use Codex to develop production-oriented agentic AI applications that can retrieve information, use tools, complete multi-step tasks, and support real-world business processes.