No. 394 of 500 No. 45 in Generative AI & LLMs

Generative AI with LangChain

Build production-ready LLM applications and advanced agents using Python, LangChain, and LangGraph

By Ben Auffarth and Leonid Kuligin

Published
May 23, 2025
Print length
476 pages
Reading level
Intermediate
ISBN-13
978-1837022014
Publisher
Packt Publishing
Written for
Python developers building retrieval-augmented and agentic LLM applications with LangChain and LangGraph.

Synopsis

Ben Auffarth's 2023 book was an early full-length guide to LangChain, and the 2025 second edition, rewritten with Google Cloud engineer Leonid Kuligin, shifts its focus from simple chains to agentic systems built for production. It assumes working Python knowledge and targets developers who have already called an LLM API. Across roughly 480 pages it moves from the history of language models and first steps with LangChain to LangGraph workflows, retrieval-augmented generation pipelines with hybrid search and re-ranking, tool-using agents and multi-agent architectures, with chapters on agents that write code and analyze data.

The later chapters address the problems that stall many generative AI projects: evaluating and testing LLM applications, deploying them, and monitoring them through observability tooling. Each chapter pairs explanation with Python code in a companion GitHub repository, which the authors keep current; the printed edition targets LangChain 0.3, and a separate branch migrates the examples to LangChain 1.0. Readers come away able to design, test and ship agent applications in the LangChain ecosystem, with an understanding of the trade-offs among retrieval strategies, reasoning patterns and agent topologies.

Scorecard

Editorial score

How it scored

Editorial score: 62.8 / 1005 weighted criteria

  • Influence30%

    16.5/30
  • Authority25%

    16.3/25
  • Practical value20%

    15.0/20
  • Currency15%

    9.0/15
  • Readability10%

    6.0/10
  • Editorial score

    62.8/100

How the rank was computed

Rank No. 394 of 500 comes from a composite of 50.1: observable signals 21.8 of 55 points plus editorial review 28.3 of 45 points.

All 11 inputs
InputPoints
Observable signals21.8/55
Reach (Amazon ratings volume)6.9/16.5
Consensus across best-of lists5.5/13.8
Awards and bestseller evidence0.0/8.3
Reader rating, sample-adjusted5.6/8.3
Amazon sales rank1.0/5.5
Recency2.8/2.8
Editorial review28.3/45
Author authority7.3/11.3
Influence7.4/13.5
Practical value6.8/9.0
Readability2.7/4.5
Currency4.1/6.8

Each criterion is scored from 1 to 10 and weighted, and the points add up to the editorial score. The rank combines observable signals (55 percent) with this score (45 percent). Read the methodology.

Verdict

A thorough, code-first route from LangChain prototypes to production LangGraph agents, best read alongside the repository's updated v1.0 branch.

Strengths

  • Covers evaluation, deployment and observability, not just prototypes
  • Runnable code in an actively maintained repository
  • Co-author maintains Google Cloud's LangChain integrations

Keep in mind

Printed code targets LangChain 0.3, so readers should use the repository's updated branch; developers on other frameworks must translate LangChain-specific code themselves.

Key takeaways

  1. LangGraph's graph-based workflows give developers explicit control over state, branching and loops, which simple LangChain chains cannot provide for complex agents.
  2. Retrieval-augmented generation quality depends on deliberate choices about chunking, embeddings, hybrid search and re-ranking, each of which needs testing.
  3. Multi-agent setups divide complex tasks among specialized agents, and LangGraph provides the structure to coordinate their handoffs and shared state.
  4. Moving from prototype to production requires systematic evaluation, testing and observability, because LLM applications fail in ways traditional unit tests miss.

Notable facts

About the author

Ben Auffarth and Leonid Kuligin

Ben Auffarth is a full-stack data scientist with more than 15 years of experience and a PhD in computational and cognitive neuroscience; he co-founded Data Science Speakers in London and is its former president. Co-author Leonid Kuligin is a staff AI engineer at Google Cloud, a key maintainer of Google Cloud's LangChain integrations and a visiting lecturer at CDTM in Munich.

Questions

What is Generative AI with LangChain about?
It teaches Python developers to build LLM applications with LangChain and LangGraph, moving from basic chains to retrieval-augmented generation, tool-using agents and multi-agent systems. The 2025 second edition includes chapters on coding and data-analysis agents, evaluation and testing, and production deployment with observability, all backed by code in a companion GitHub repository.
Is Generative AI with LangChain still current in 2026?
Partly. The concepts, including LangGraph workflows, RAG design, agent patterns and evaluation, remain relevant, but the printed code targets LangChain 0.3. The authors maintain a repository branch updated for LangChain 1.0, so readers should run that version instead of copying from the page. Teams on other frameworks will still find the architecture chapters useful.
Who should read Generative AI with LangChain?
Python developers and machine learning engineers who are building chatbots, retrieval systems or agents and want a structured path to production will benefit most. The book assumes programming experience and some familiarity with LLM APIs. Managers and beginners seeking a conceptual introduction to generative AI will find it too code-heavy and should start with a less technical title.

Sources

As an Amazon Associate, Best AI Books earns from qualifying purchases. Best AI Books is sponsored by Iternal.ai, whose founder and CEO, John Byron Hanby IV, wrote The AI Strategy Blueprint, our #1 ranked book. How we rank books.