Essential SE Foundations for Agentic Software Engineering
Not a book about prompt engineering. A book about applying rigorous software engineering principles when AI agents are in the loop.
Who this is for
Software engineering students
Learning SE foundations in a world where agents are part of the toolkit from day one.
Mid-career engineers upskilling
Transitioning from AI-assisted coding to deliberate agentic engineering practices.
Educators building curricula
Designing courses and assessments that develop AI-native capability, not just familiarity.
Table of contents
What changes when agents join the team
How the SDLC shifts when AI agents can plan, code, and execute — not just suggest.
Requirements and intent in an agentic world
Specifying outcomes, constraints, and acceptance criteria for agent-ready tasks.
Designing agent-ready tasks
Decomposing work into small, reviewable, low-risk units for agents to execute.
Code review and verification with AI outputs
Reviewing AI-generated changes with the rigour of a human PR review.
Testing strategies for generated code
How to verify AI output — and how AI can help you verify its own output.
Security, privacy, and agentic risk
Threat models, prompt injection, secrets leakage, and agentic attack surfaces.
Architecture and technical debt under AI acceleration
Keeping design quality when agents can generate large amounts of code quickly.
Guardrails, governance, and human approval
Building permission models, audit trails, and approval gates before scaling autonomy.
Learning and assessment in agentic SE education
How students and educators work with AI-native tools while developing genuine capability.
The AI-native engineer: professional practice
Trust calibration, career development, and the long-term trajectory of the discipline.
Educator resources
The book is designed for classroom use. Instructor materials, slide decks, assessments, and workshop activities are available through EdCompass.
Go to EdCompass