Chapter 12: AI and the Architect's New Decisions

What AI integration means for system architecture: LLM APIs, retrieval-augmented generation, model deployment, data pipelines for AI, and the reliability and safety considerations that come with non-deterministic components.

What this chapter covers

LLM API integration patterns and latency management
Retrieval-Augmented Generation (RAG) architecture
Vector databases and semantic search
ML model deployment and serving infrastructure
Data pipelines for AI/ML systems
Non-determinism and testing AI components
AI safety and guardrails in production systems
Cost architecture for LLM-powered features

Read Chapter 12 in Full

Available on LeanPub (PDF, ePub, MOBI) and Amazon (Kindle and Paperback).

Buy on LeanPub All 28 Chapters →

← Ch 11: Observability: Designing Systems You Can Understand

Ch 13: Cloud Architecture: Building for the Platform →