Artículos relacionados a UNDERSTANDING AND BUILDING AI AGENTS: A Practical Guide...

UNDERSTANDING AND BUILDING AI AGENTS: A Practical Guide to Designing, Developing, and Deploying Intelligent Autonomous Systems - Tapa blanda

Thornqvist, Leif

 
9798166098153: UNDERSTANDING AND BUILDING AI AGENTS: A Practical Guide to Designing, Developing, and Deploying Intelligent Autonomous Systems

Sinopsis

Building an AI agent is not just about getting a model to produce a plan. It is about designing a system that can act within defined limits—and show whether its actions worked.

An agent may select the right tool but supply invalid arguments, rely on outdated information, or report success before an operation is confirmed. These problems become harder when a workflow involves retrieval, multiple steps, outside services, and human approvals. The challenge is to connect intelligent decision-making with testable software engineering.
Understanding and Building AI Agents provides a practical, structured introduction to designing, developing, and evaluating goal-directed software agents. It explains how models, controllers, tools, memory, planning, and validation fit together so readers can assess not only what an agent proposes, but what it is authorized to execute.

  • Understand agent fundamentals, autonomy boundaries, observations, goals, actions, and feedback.
  • Design modular agent architectures with state, structured outputs, tools, API contracts, and explicit permissions.
  • Explore knowledge retrieval, embeddings, memory, and grounded response generation.
  • Organize tasks with planning, workflow automation, retries, recovery, and stopping conditions.
  • Compare single-agent and multi-agent approaches, including coordination and shared state.
  • Evaluate results through test cases, measurable criteria, traces, security checks, and human review.
  • Connect the pieces through deployment guidance, worked examples, exercises, and an integrated capstone.
Across twelve chapters, the book moves from foundational ideas and Python-oriented development to real-world architectural trade-offs. Its worked examples, diagrams, practice problems, selected answers, glossary, and troubleshooting checklist support both systematic study and later reference.
It is intended for technically curious newcomers, software developers, and professionals evaluating where agent-based methods belong. The book does not claim that an example is automatically secure or production-ready; it emphasizes testing, adaptation, and accountable operation.

Get your copy and develop a clearer engineering approach to agents that can make decisions, use tools, and remain subject to meaningful controls.

"Sinopsis" puede pertenecer a otra edición de este libro.