
AI-Native Engineering
Building Software with AI Agents
A practical guide to working with AI agents throughout the software development process.
Learn how to give agents the context they need, turn requirements into actionable tasks, verify their changes, and keep your codebase understandable as your use of AI grows.
About the book
AI agents can explore a codebase, write code, run tests, and help with reviews. Making good use of them requires more than a well-written prompt: engineers still need to make decisions, understand the changes, and check whether the result solves the right problem.
AI-Native Engineering explores practical ways to work with agents across the development process, from the first conversation about a task to implementation, verification, and team adoption.
The book uses iOS development examples to illustrate these ideas, while the practices can be applied across platforms and technology stacks.
What you’ll learn
- Give AI agents useful context about an existing project.
- Turn requirements into clear, manageable tasks.
- Use skills, tools, and MCP to extend an agent’s capabilities.
- Organize development workflows and review agent-generated changes.
- Verify behavior through testing, visual checks, and other evidence.
- Work with multiple agents and manage longer tasks.
- Evaluate agent performance and consider security risks.
- Keep a codebase understandable as AI-assisted development grows.
Inside the book
The book is organized into four parts:
Part I — Agentic Development Foundations
The fundamentals of working with agents, providing context, defining tasks, and organizing a development workflow.
Part II — Verification and Quality
Testing, visual verification, debugging, and code review practices that help you assess agent-generated changes.
Part III — Advanced Agentic Workflows
Architecture migration, multi-agent development, long-running tasks, and agent security.
Part IV — Evaluation and Engineering Responsibility
Agent evaluations, cognitive debt, and practices for introducing AI agents into a team.
Who this book is for
This book is for software engineers, technical leads, and engineering teams who want to use AI agents in their work while maintaining a clear understanding of the software they build.
The examples draw on iOS development, but you don’t need to be an iOS developer to apply the broader ideas.
