
The bookLaunching soon
AI Engineering, Measured
From Developer to AI Engineer, Backed by Reproducible Runs
by Hassan Mansour
A book for working developers who know how to ship software and want real tradeoffs about building with AI, not a sales pitch. It covers retrieval, agents, evaluation, security, observability and cost.
Every number in this book came from a run.
There is no order page yet. Join the waitlist below to hear once when it opens.
In numbers
- 15 chapters
- in 3 parts, plus a closing chapter
- 143 measured findings
- counted from the book's own record
- 4 appendices
- The book is written in English.
What is inside
Part 1The Mental Shift
- 1Why AI Development Feels Different
- 2Picking Your Stack Without Getting Paralyzed
- 3Structured Output: Getting Data Back, Not Text
- 4Your First Agent: A Loop, a Tool, and a Stop Condition
Part 2Core Building Blocks
- 5RAG: When to Retrieve, and How to Tell It Failed
- 6Vector Storage: An Index That Outlives the Process
- 7Conversation State: What to Keep and What to Drop
- 8Multi-Agent Systems: When to Split, and When Not To
- 9MCP: Using Tools You Did Not Write
Part 3Production Reality
- 10Evaluation: Catching Regressions in a System That Won't Repeat
- 11Security: Bounding What the Model Can Do
- 12Observability: Tracing What Actually Happened
- 13Cost: What It Spends, and How to Spend Less
- 14Streaming, Async, and Jobs: Handling the Wait
15Closing chapterYour Own System: The First Three Moves
Waitlist
Waitlist opens soon
There is no form here yet, so this page collects nothing about you.