+- MKLab (https://mklab.gr)
+-- Forum: [INDEX] (https://mklab.gr/forumdisplay.php?fid=1)
+--- Forum: ARTFICIAL INTELLIGENCE (AI) (https://mklab.gr/forumdisplay.php?fid=5)
+---- Forum: BOOKS (https://mklab.gr/forumdisplay.php?fid=32)
+----- Forum: NEW PUBLICATIONS (https://mklab.gr/forumdisplay.php?fid=110)
+----- Thread: An Illustrated Guide to AI Agents [Grootendorst] (/showthread.php?tid=1973)
An Illustrated Guide to AI Agents [Grootendorst] - mklabgr - 09-15-2026
An Illustrated Guide to AI Agents by Maarten Grootendorst and Jay Alammar is a practical, intuition-first introduction to how modern AI agents work. Rather than treating an agent as simply an LLM with a prompt, the book explains an agent as a system that combines an LLM with capabilities such as reasoning, memory, tools, planning, reflection, and interaction with its environment.
It begins with the foundations of LLMs—including transformers, self-attention, mixture-of-experts models, and test-time reasoning—and then builds toward more sophisticated agent architectures. The authors use extensive visual explanations and a progressively developed “TinyAgent” to make the underlying mechanisms concrete.
The later chapters focus on the components that make agents genuinely autonomous: short- and long-term memory, retrieval-augmented generation, tool calling, tool learning, Model Context Protocol (MCP), planning, task decomposition, reflection, and evaluation. It also covers multimodal agents, coding agents, multi-agent collaboration, and techniques such as distillation, quantization, supervised fine-tuning, and reinforcement learning.
Overall, the book is positioned as both a conceptual guide and a hands-on technical reference for understanding and building agentic AI systems, particularly useful for readers who want to move beyond using LLMs toward actually designing AI agents.
BOOK