Agentic AI for Serious Engineers¶
A practical field guide to building reliable, evaluable, and production-grade agent systems
Most agentic AI material teaches you how to make an impressive demo. This book teaches engineers how to build agent systems that survive real-world constraints.
This is written for backend engineers, platform engineers, and staff+ engineers who understand distributed systems, reliability, and the cost of complexity -- but are relatively new to AI. You do not need a machine learning background. You need the judgment to know when AI helps and when it makes things worse.
What makes this different is the engineering-first approach. Every concept comes with failure modes, tradeoffs against alternatives, and evaluation criteria. There is no vibes-based testing, no "just prompt it better" advice, and no framework worship. The goal is transferable understanding: principles that hold regardless of which SDK ships next month.
Chapters¶
| # | Chapter | What you learn |
|---|---|---|
| 1 | What "Agentic" Actually Means | Precise vocabulary: LLM app vs workflow vs agent vs multi-agent |
| 2 | Tools, Context, and the Agent Loop | Building blocks: tool registry, context engineering, observe-think-act |
| 3 | Workflow-First, Agent-Second | The most important architectural decision |
| 4 | Multi-Agent Systems Without Theater | When multiple agents help and when they are complexity theater |
| 5 | Human-in-the-Loop as Architecture | Approval gates, escalation, and auditability |
| 6 | Evaluating and Hardening Agent Systems | Eval harnesses, tracing, reliability, cost, security |
| 7 | When Not to Use Agents | The signature chapter -- judgment over hype |
Learning Paths¶
| Path | Chapters | Time |
|---|---|---|
| Fast Engineer | 1, 2, 7 | ~1 hour |
| Full Mastery | 1 through 7 in order | ~4 hours |
| Enterprise Architect | 1, 3, 5, 6, 7 | ~2.5 hours |
Projects¶
Two end-to-end systems built incrementally through the chapters:
- Document Intelligence Agent -- Ingest documents, retrieve evidence, answer with citations, escalate on uncertainty
- Incident Runbook Agent -- Inspect signals, search runbooks, propose remediation, request human approval
This book is free to read online. Licensed under CC BY-NC-SA 4.0 -- share and adapt with credit, non-commercial use only.