Generative & Agentic AI
Agentic AI and MCP Integration
G02 · Practitioner
Entry requirements
G01; Python functions and APIs
What 36 hours means
- Lessons
- 24 h
- Capstone
- 8 h
- Assessment
- 2 h
- Remediation
- 2 h
Syllabus
Four modules, twelve lessons. Every lesson states what it covers and the evidence you produce for it.
Module 1
Agent design
Agents and bounded workflows · State and planning · Tool contracts
G02Show lessonsHide lessons
G02.01 · 120 min
Agents and bounded workflows
Decide whether an agent or deterministic workflow suits a task.
Evidence: A design states allowed actions, termination and human responsibility.
G02.02 · 120 min
State and planning
Represent task state and separate proposed steps from executed steps.
Evidence: A state trace prevents repetition after a completed action.
G02.03 · 120 min
Tool contracts
Module 2
Tools and MCP
MCP concepts · MCP server exercise · Authentication and scope
G02Show lessonsHide lessons
G02.04 · 120 min
MCP concepts
Map client, server, tools and resources in a controlled integration.
Evidence:
Module 3
Control and recovery
Memory and context · Human approval · Retries and idempotency
G02Show lessonsHide lessons
G02.07 · 120 min
Memory and context
Choose task-local state and redact persistent notes.
Evidence:
Module 4
Evaluation
Agent evaluation · Multi-agent boundaries · Observability and release
G02Show lessonsHide lessons
G02.10 · 120 min
Agent evaluation
Assess tool selection, argument accuracy and final outcomes.
Evidence:
Capstone
Build a delivery-exception assistant using mock customer, locker and playbook tools. Require approval for external actions and evaluate action correctness, privacy and recovery.
Lab environment
AI API sandbox
Lab infrastructure is not yet provisioned, so this describes the environment the course is designed for rather than one you can launch today.