Agent foundations and tools
Explain the agent loop and build validated, tested tools that a model can call.
AI Agents and Automation Builder
Build AI agents that use tools, follow approval rules and recover from failures, then test them before they touch real systems. Proposed programme; academic approval and batch confirmation pending.
Explore what you’ll learn.
Build, test and deploy tool-using AI agents with human approval, injection defences and budget limits.
Proposed curriculum · academic approval pendingSYLLABUS OVERVIEW
A structured route through the subject
Core learning areas at a glance. Detailed units and teaching plans are in the private learning workspace.
Planning, memory and approved knowledge
Build multi-step agents that plan within limits and answer only from approved, cited sources.
Human approval and safety
Gate consequential actions behind human approval and defend against prompt injection.
Evaluation, reliability and cost
Measure agent quality with a harness and control failures, retries and spend.
Deployed agent project
Deploy, evaluate and present a sandboxed agent with clear limitations.
Entry requirements and learning outcomes
Comfortable writing Python functions, classes and tests Calling REST APIs and handling JSON responses Basic Git and command-line use Access to an LLM API account approved by the institute, with spending limits set
- Explain how a tool-using agent loop works and where the application, not the model, executes actions.
- Design, validate and test tools with JSON schemas over REST APIs.
- Build agents that plan multi-step tasks within step limits and answer from approved knowledge with citations.
- Place human approval gates and least-privilege controls on consequential actions.
- Test agents against prompt injection and evaluate them with a repeatable harness.
YOUR PRACTICAL PORTFOLIO
Customer support agent with approved knowledge and refund approval
Build, check and explain your work with trainer feedback. Certificate and assessment terms are confirmed by the institute.
Continue in your learning workspace
Open detailed curriculum ↗Detailed syllabus and weekly plans are available to authorised staff and learners enrolled in this course.