Product guide

Building Agent-Powered Applications: Generative AI, RAG, Fine-Tuning, and Agents

Agents: Building Agent-Powered Applications is a Packt book on engineering generative AI systems from initial model concepts through production-oriented workflows. It covers prompting, RAG, fine-tuning, agent design, evaluation, and deployment, with practical trade-offs for readers who already have programming and software engineering fundamentals.

At a glance

Moves from AI, NLP, embeddings, transformers, and LLM behavior into prompting, RAG, fine-tuning, and agent workflows.
Covers agent tools, memory, planning, orchestration, and human-in-the-loop controls.
Includes evaluation topics such as offline and online testing, task-specific metrics, LLM-as-a-judge methods, and responsible AI checks.
The listing says print or Kindle purchases include a free PDF eBook. Access requires email sign-up and proof of purchase.

Price

$33.99Save 32%

Reference price: ~~$49.99~~

Price checked: Nov 14, 2026

View offer Ahead of Black FridayExplore the selection

Pros and cons

Pros

Combines model fundamentals, application techniques, agent design, and evaluation in one end-to-end engineering-focused book.
Includes guidance on choosing between prompting, retrieval-augmented generation, and fine-tuning rather than treating them as unrelated methods.
The stated audience includes engineers transitioning from traditional software or classical machine learning, and says deep learning or LLM training experience is not required.

Technical specifications

8 specifications

Book details
Publisher	Packt Publishing
Publication date	April 30, 2026
Language	English
Print length	490 pages
ISBN-10	1807605175
ISBN-13	978-1807605179

Dimensions and installation
Weight listed	1.84 pounds
Dimensions	7.5 x 1.11 x 9.25 inches

What to know before buying

The listing describes a free PDF and access to Packt's Reader, but the PDF and Reader access require email sign-up and proof of purchase.

Who is it for?

A good fit for

Software engineers, AI engineers, data scientists, applied AI practitioners, and technical leads building LLM applications.
Engineering-focused product managers who want an overview of production-oriented generative AI architectures and trade-offs.
Programmers comfortable with Python or a similar language, APIs, data structures, and integration who are moving into generative AI.

Less suited to

Readers seeking a programming-free introduction, since the listing expects programming and core software engineering familiarity.
Readers looking specifically for deep learning or LLM training instruction, which is not presented as a required background or central focus.

Customer reviews on Amazon

4.4 / 5 · 10 ratings

5 stars74%74%
4 stars11%11%
3 stars0%0%
2 stars15%15%
1 star0%0%

Aggregate ratings lean positive, with most ratings in the five-star category, but a smaller two-star share means the overall picture is not uniformly favorable. The sample is limited, and no written feedback is supplied here to explain either the positive or critical ratings. These figures cannot establish how useful the book's coverage of RAG, agents, and production evaluation will be for a particular reader.

Frequently asked questions

What topics does Building Agent-Powered Applications cover?

The listed topics span AI and NLP fundamentals, LLMs, prompting, language tasks, RAG, fine-tuning, agent architecture and construction, and evaluation of LLM applications and agents.

Who is this book intended for?

The listing targets engineers, data scientists, applied AI practitioners, technical leads, and engineering-focused product managers. It expects programming ability and familiarity with software engineering concepts such as APIs, data structures, and integration.

Does the book include a digital version?

The listing says purchase of the print or Kindle book includes a free PDF eBook. It requires email sign-up and proof of purchase, and also mentions access to Packt's Reader.

Is prior deep learning experience required?

The listing says prior deep learning or LLM training experience is not required, while programming and core software engineering concepts are expected.