Agentic Design Patterns (docs.google.com)

Antonio Gulli’s Agentic Design Patterns is a hands-on, 424-page practitioner’s guide that systematically catalogs reusable patterns for building “agentic” AI systems. Organized into four parts and a hefty appendix, the book covers foundational patterns (prompt chaining, routing, parallelization, reflection, tool use, planning, multi-agent), system-level concerns (memory management, learning/adaptation, Model Context Protocols, goal setting/monitoring), robustness and human collaboration (exception recovery, human-in-the-loop, retrieval-augmented generation), and advanced topics (inter-agent communication, resource-aware optimization, reasoning techniques, guardrails, evaluation and prioritization). Each chapter includes runnable code and practical examples, plus online supplements (AgentSpace tutorials, CLI agents) and a generated index demonstrating an agentic reasoning step. All royalties will be donated to Save the Children. For the AI/ML community this is significant because it bridges conceptual agent research and production engineering: you get concrete modules and design patterns that enable composability, scalable coordination, stateful memory, adaptive learning, and safety/monitoring strategies. Key technical takeaways include standardized Model Context Protocols for context handling, patterns for multi-agent A2A communication and prioritization, practical RAG and memory-management recipes, and evaluation/guardrail templates to operationalize reliability and compliance. In short, the book functions as a pattern language for constructing practical, auditable, and resource-aware intelligent agents.