book

AI Agents in Action

Name: AI Agents in Action
Author: Micheal Lanham
ISBN: 9781633436343

by Micheal Lanham

February 2025

Intermediate to advanced

344 pages

9h 2m

English

Manning Publications

Read now

Unlock full access

AI Agents in Action
copyright
dedication
contents
preface
acknowledgments
about this book
Who should read this bookHow this book is organized: A road mapAbout the codeliveBook discussion forum
about the author
about the cover illustration
1 Introduction to agents and their world
1.1 Defining agents1.2 Understanding the component systems of an agent1.3 Examining the rise of the agent era: Why agents?1.4 Peeling back the AI interface1.5 Navigating the agent landscapeSummary

2 Harnessing the power of large language models
2.1 Mastering the OpenAI API2.1.1 Connecting to the chat completions model2.1.2 Understanding the request and response2.2 Exploring open source LLMs with LM Studio2.2.1 Installing and running LM Studio2.2.2 Serving an LLM locally with LM Studio2.3 Prompting LLMs with prompt engineering2.3.1 Creating detailed queries2.3.2 Adopting personas2.3.3 Using delimiters2.3.4 Specifying steps2.3.5 Providing examples2.3.6 Specifying output length2.4 Choosing the optimal LLM for your specific needs2.5 ExercisesSummary
3 Engaging GPT assistants
3.1 Exploring GPT assistants through ChatGPT3.2 Building a GPT that can do data science3.3 Customizing a GPT and adding custom actions3.3.1 Creating an assistant to build an assistant3.3.2 Connecting the custom action to an assistant3.4 Extending an assistant’s knowledge using file uploads3.4.1 Building the Calculus Made Easy GPT3.4.2 Knowledge search and more with file uploads3.5 Publishing your GPT3.5.1 Expensive GPT assistants3.5.2 Understanding the economics of GPTs3.5.3 Releasing the GPT3.6 ExercisesSummary
4 Exploring multi-agent systems
4.1 Introducing multi-agent systems with AutoGen Studio4.1.1 Installing and using AutoGen Studio4.1.2 Adding skills in AutoGen Studio4.2 Exploring AutoGen4.2.1 Installing and consuming AutoGen4.2.2 Enhancing code output with agent critics4.2.3 Understanding the AutoGen cache4.3 Group chat with agents and AutoGen4.4 Building an agent crew with CrewAI4.4.1 Creating a jokester crew of CrewAI agents4.4.2 Observing agents working with AgentOps4.5 Revisiting coding agents with CrewAI4.6 ExercisesSummary
5 Empowering agents with actions
5.1 Defining agent actions5.2 Executing OpenAI functions5.2.1 Adding functions to LLM API calls5.2.2 Actioning function calls5.3 Introducing Semantic Kernel5.3.1 Getting started with SK semantic functions5.3.2 Semantic functions and context variables5.4 Synergizing semantic and native functions5.4.1 Creating and registering a semantic skill/plugin5.4.2 Applying native functions5.4.3 Embedding native functions within semantic functions5.5 Semantic Kernel as an interactive service agent5.5.1 Building a semantic GPT interface5.5.2 Testing semantic services5.5.3 Interactive chat with the semantic service layer5.6 Thinking semantically when writing semantic services5.7 ExercisesSummary
6 Building autonomous assistants
6.1 Introducing behavior trees6.1.1 Understanding behavior tree execution6.1.2 Deciding on behavior trees6.1.3 Running behavior trees with Python and py_trees6.2 Exploring the GPT Assistants Playground6.2.1 Installing and running the Playground6.2.2 Using and building custom actions6.2.3 Installing the assistants database6.2.4 Getting an assistant to run code locally6.2.5 Investigating the assistant process through logs6.3 Introducing agentic behavior trees6.3.1 Managing assistants with assistants6.3.2 Building a coding challenge ABT6.3.3 Conversational AI systems vs. other methods6.3.4 Posting YouTube videos to X6.3.5 Required X setup6.4 Building conversational autonomous multi-agents6.5 Building ABTs with back chaining6.6 ExercisesSummary
7 Assembling and using an agent platform
7.1 Introducing Nexus, not just another agent platform7.1.1 Running Nexus7.1.2 Developing Nexus7.2 Introducing Streamlit for chat application development7.2.1 Building a Streamlit chat application7.2.2 Creating a streaming chat application7.3 Developing profiles and personas for agents7.4 Powering the agent and understanding the agent engine7.5 Giving an agent actions and tools7.6 ExercisesSummary
8 Understanding agent memory and knowledge
8.1 Understanding retrieval in AI applications8.2 The basics of retrieval augmented generation (RAG)8.3 Delving into semantic search and document indexing8.3.1 Applying vector similarity search8.3.2 Vector databases and similarity search8.3.3 Demystifying document embeddings8.3.4 Querying document embeddings from Chroma8.4 Constructing RAG with LangChain8.4.1 Splitting and loading documents with LangChain8.4.2 Splitting documents by token with LangChain8.5 Applying RAG to building agent knowledge8.6 Implementing memory in agentic systems8.6.1 Consuming memory stores in Nexus8.6.2 Semantic memory and applications to semantic, episodic, and procedural memory8.7 Understanding memory and knowledge compression8.8 ExercisesSummary
9 Mastering agent prompts with prompt flow
9.1 Why we need systematic prompt engineering9.2 Understanding agent profiles and personas9.3 Setting up your first prompt flow9.3.1 Getting started9.3.2 Creating profiles with Jinja2 templates9.3.3 Deploying a prompt flow API9.4 Evaluating profiles: Rubrics and grounding9.5 Understanding rubrics and grounding9.6 Grounding evaluation with an LLM profile9.7 Comparing profiles: Getting the perfect profile9.7.1 Parsing the LLM evaluation output9.7.2 Running batch processing in prompt flow9.7.3 Creating an evaluation flow for grounding9.7.4 ExercisesSummary
10 Agent reasoning and evaluation
10.1 Understanding direct solution prompting10.1.1 Question-and-answer prompting10.1.2 Implementing few-shot prompting10.1.3 Extracting generalities with zero-shot prompting10.2 Reasoning in prompt engineering10.2.1 Chain of thought prompting10.2.2 Zero-shot CoT prompting10.2.3 Step by step with prompt chaining10.3 Employing evaluation for consistent solutions10.3.1 Evaluating self-consistency prompting10.3.2 Evaluating tree of thought prompting10.4 ExercisesSummary
11 Agent planning and feedback
11.1 Planning: The essential tool for all agents/assistants11.2 Understanding the sequential planning process11.3 Building a sequential planner11.4 Reviewing a stepwise planner: OpenAI Strawberry11.5 Applying planning, reasoning, evaluation, and feedback to assistant and agentic systems11.5.1 Application of assistant/agentic planning11.5.2 Application of assistant/agentic reasoning11.5.3 Application of evaluation to agentic systems11.5.4 Application of feedback to agentic/assistant applications11.6 ExercisesSummary
appendix A Accessing OpenAI large language models
A.1 Accessing OpenAI accounts and keysA.2 Azure OpenAI Studio, keys, and deployments
appendix B Python development environment
B.1 Downloading the source codeB.2 Installing PythonB.3 Installing VS CodeB.4 Installing VS Code Python extensionsB.5 Creating a new Python environment with VS CodeB.6 Using VS Code Dev Containers (Docker)
index

Content preview from AI Agents in Action

2 Harnessing the power of large language models

This chapter covers

Understanding the basics of LLMs
Connecting to and consuming the OpenAI API
Exploring and using open source LLMs with LM Studio
Prompting LLMs with prompt engineering
Choosing the optimal LLM for your specific needs

The term large language models (LLMs) has now become a ubiquitous descriptor of a form of AI. These LLMs have been developed using generative pretrained transformers (GPTs). While other architectures also power LLMs, the GPT form is currently the most successful.

LLMs and GPTs are generative models, which means they are trained to generate rather than predict or classify content. To illustrate this further, consider figure 2.1, which shows the difference ...

Become an O’Reilly member and get unlimited access to this title plus top books and audiobooks from O’Reilly and nearly 200 top publishers, thousands of courses curated by job role, 150+ live events each month,
and much more.

Read now

Unlock full access

More than 5,000 organizations count on O’Reilly

O’Reilly covers everything we've got, with content to help us build a world-class technology community, upgrade the capabilities and competencies of our teams, and improve overall team performance as well as their engagement.

Julian F.

Head of Cybersecurity

I wanted to learn C and C++, but it didn't click for me until I picked up an O'Reilly book. When I went on the O’Reilly platform, I was astonished to find all the books there, plus live events and sandboxes so you could play around with the technology.

Addison B.

Field Engineer

I’ve been on the O’Reilly platform for more than eight years. I use a couple of learning platforms, but I'm on O'Reilly more than anybody else. When you're there, you start learning. I'm never disappointed.

Amir M.

Data Platform Tech Lead

I'm always learning. So when I got on to O'Reilly, I was like a kid in a candy store. There are playlists. There are answers. There's on-demand training. It's worth its weight in gold, in terms of what it allows me to do.

Mark W.

Embedded Software Engineer

Publisher Resources

ISBN: 9781633436343Publisher Support Publisher Website

Cloud Computing

Data Engineering

Data Science

AI & ML

Programming Languages

Software Architecture

IT/Ops

Security

Design

Business

Soft Skills

AI Agents in Action

by Micheal Lanham