Agentic Design Patterns

Software engineering has long relied on design patterns to give developers a shared vocabulary for solving recurring problems, and Agentic Design Patterns applies that same approach to the comparatively new and chaotic world of AI agents. Antonio Gullí identifies recurring challenges in building agentic systems — how agents plan, how they use tools, how they handle memory and state, how multiple agents coordinate — and organizes solutions to each into clear, reusable patterns that engineers can apply directly to their own projects. This structural approach brings welcome order to a field that’s often documented through scattered blog posts and shifting best practices, giving readers something closer to a stable reference. At 472 pages, the book has room to develop each pattern with real depth, including the reasoning behind why certain approaches work better than others in specific contexts. For engineers building agentic systems who want a more principled, pattern-based foundation rather than ad hoc trial and error, this fills a genuine gap in the current literature.

Author:
Antonio Gullí