Building AI Agents for Complex Tasks

位教师：Hurix Digital

访问权限由 Coursera Learning Team 提供

3个模块

深入了解一个主题并学习基础知识。

中级等级

推荐体验

4 小时完成

灵活的计划

自行安排学习进度

3个模块

深入了解一个主题并学习基础知识。

中级等级

推荐体验

4 小时完成

灵活的计划

自行安排学习进度

您将获得的技能

要了解的详细信息

可分享的证书

添加到您的领英档案

作业

7 任务¹

AI 评分请参见免责声明

授课语言：英语（English）

了解顶级公司的员工如何掌握热门技能

了解关于 Coursera for Business 的更多信息

Petrobras, TATA, Danone, Capgemini, P&G 和 L'Oreal 的徽标

该课程共有3个模块

Building AI Agents for Complex Tasks is an intermediate-level course designed to equip learners with the skills to design, build, and evaluate intelligent agents that operate autonomously across dynamic, multi-step environments. Moving beyond simple chatbot flows, this course introduces learners to agent architectures that perceive context, make decisions, integrate tools, and recover from failure.

Through hands-on labs, interactive video walkthroughs, and real-world case studies—including Alexa, BabyAGI, and AlphaCode—learners will explore agent types, design patterns, tool orchestration, memory management, and behavior evaluation. They'll gain practical experience using modern frameworks like LangChain and Rasa to construct agents for use cases such as research automation, virtual assistants, and decision-making bots. By the end of the course, learners will have built and tested their own intelligent agent and developed the foundational skills to implement agent-based AI systems that can adapt, reason, and act in real-world applications.

This foundational lesson introduces what AI agents are and how they differ from traditional software. Learners will explore agent-environment interactions, the concept of perception, and how various types of agents—reactive, deliberative, and hybrid—handle decision-making. Through real-world examples like smart assistants and warehouse robots, learners will classify agent types and determine where each model excels or breaks down.

涵盖的内容

4个视频1篇阅读材料1个作业

4个视频总计18分钟

Introduction and Welcome 2分钟
How AI Agents Perceive the World Around Them 7分钟
Reactive vs. Deliberative Agents in the Real World 5分钟
Rasa vs. AutoGPT – Choosing the Right Agent Model 4分钟

1篇阅读材料总计6分钟

Welcome to the Course: Course Overview 6分钟

1个作业总计10分钟

HOL: Practice Classifying Agent Types in Real-World Use Cases 10分钟

This lesson moves from theory to implementation. Learners will construct intelligent agents that integrate inputs (perception), structured reasoning (decision loops), and output (action). They'll explore core modules such as memory, planning chains, and tool execution in LangChain and Rasa. Real-world examples like Alexa’s task-based updates and LangChain agents with tools will help frame the technical walkthroughs.

涵盖的内容

3个视频1篇阅读材料2个作业

3个视频总计15分钟

How Intelligent Agents Turn Input into Impact 5分钟
Designing Action Loops and Memory Modules 5分钟
Alexa Multi-Intent Flow and Long Chain Tool Use 5分钟

1篇阅读材料总计8分钟

LangChain and Rasa: Agent Workflows and APIs 8分钟

2个作业总计20分钟

HOL: Build a Simple Multi-Step Agent Using LangChain or Rasa 10分钟
HOL: Intelligent Agent Lab 10分钟

In the final lesson, learners will focus on evaluating how agents perform in realistic, changing environments. They'll explore testing strategies, interpret edge-case behaviors, and fine-tune agents using logs, performance feedback, and outcome tracking. Examples such as AlphaCode’s reasoning iterations and BabyAGI’s task queue refinement will help frame the concepts. This lesson culminates in the Capstone project, where learners will apply everything they've learned to design and deliver an intelligent, goal-driven agent.