The Weekly Roundup!

Save up to 45% on your entire purchase.

Expires midnight PT on May 17th. Only at manning.com.

Note: Final discount will be applied in cart. For products already on sale, the best available discount will be applied.

You’ll begin by learning what agentic AI is, why augmented LLMs and tool use matter, and what makes these systems difficult in practice. From there, it moves step-by-step from the simplest useful unit—a single agent—to increasingly capable systems such as multi-agent crews, MCP-exposed agents, multimodal crews, complex flows, chatbot interfaces, extensible systems, and human-in-the-loop workflows.

We think Max Gfeller's perspective is a real highlight: He approaches the subject like a builder, emphasizing architecture, workflow design, and practical integration rather than hype. Unlike the fragmented picture you get from focused tutorials or short videos, you’ll finish this book with a coherent end-to-end mental model, a clear understanding of the benefits and tradeoffs, the ability to connect new concepts to your existing knowledge, and a lasting reference you can use as you move from experimentation to production! [Read more]

4 chapters of this new MEAP are available now, with more to follow soon!

This engaging book begins by establishing a solid conceptual understanding of how large language models actually work, how they manage to sound so human, and what “thinking” really means for a machine. You’ll explore the big ideas behind machine consciousness, including Alan Turing’s famous “Turing Test” that measures a machine's ability to exhibit intelligent behavior indistinguishable from a human and research into LLM “hallucinations,” with just the right balance between technical detail and AI theory. [Read more]

4 chapters of this new MEAP are available now with more to follow soon!

pipeline. It starts by deriving quantization mapping from first principles, and then builds your knowledge and skill through techniques for production-tested PTQ and QAT workflows and a fully-compressed deployment. You'll learn to apply post-training quantization to production models, run quantization-aware training using fake quantization and straight-through estimators, and handle subtle tradeoffs like activation outliers in LLMs, KV cache pressure, and sub-8-bit formats like NF4 and FP4. [Read more]

4 chapters of this new MEAP are available now with more to follow soon!

If you don’t ground your LLM, there’s a high likelihood that your GenAI app will hallucinate. Without the right context, your AI will likely provide wrong or even nonsensical results.

There’s a fix: the tag team of RAG and knowledge graphs. If you connect your data using a knowledge graph, you supply the context your LLM needs to give more nuanced, accurate responses. You can even combine structured and unstructured data.

We’ve teamed up with Neo4J to give away copies of Essential GraphRAG, a comprehensive guide on how to build a GraphRAG system from scratch. You’ll learn basic RAG concepts, advanced techniques, 

and how to evaluate accuracy for your GenAI application.

and decision-making. [Read more]

Secrets of the JavaScript Ninja, Third Edition >> Chapter 11, “Buffers and streams” >> Examine two “closer to the metal” solutions for handling the massive amounts of data that modern processing tasks demand (and that conventional JavaScript arrays just can’t handle).

Rearchitecting LLMs >> Chapter 5, “Shaping model architectures via width pruning” & Chapter 6, “Knowledge recovery through distillation” >> Examine a technique that can be used with precision to change your model’s personality—and learn why you’ll want to. Then, walk through one of the most critical stages in the model rearchitecting pipeline (hint: it's what determines whether the model can match, or even beat, the base model).

Grokking Machine Learning, Second Edition >> Chapter 9, “Splitting data by asking questions: Decision trees” & Chapter 10, “Combining building blocks to gain more power: Neural networks” >> It’s a yes-or-no question that begins the process of splitting data. Branch out to decision trees, powerful models that use a series of yes-or-no questions that help the models make predictions—and tell us a lot about our data along the way. 

Then, take a look at a type of ML model that’s so useful that it’s been given a name—and a field—all its own.

Learn AI Data Engineering in a Month of Lunches >> Chapter 11, “AI-powered web scraping” >> That’s the question you’ll ask yourself as you walk through each stage of the enrichment pipeline, learning how to identify potential pain points for when you want to scale up, and ultimately building a framework for spotting AI opportunities in any data engineering workflow.