Why It Matters for Practitioners
Working developers building AI features face a recurring decision tree: when does a better prompt suffice, when is RAG the right answer, and when does fine-tuning justify the cost? Most tutorials treat these as independent topics. "The Developer's Guide to AI" sequences them as a layered build - you ship a local LLM app first, then add prompt discipline, then retrieval, then fine-tuning judgment, then agentic tool use - so each technique gets introduced in context of what came before. That structure is the book's clearest differentiator.
The Book
"The Developer's Guide to AI: A Field Guide for the Working Developer" by Jacob Orshalick, Jerry M. Reghunadh, and Danny Thompson, published by No Starch Press, San Francisco, 2026 (320 pages, ISBN 978-1-7185-0476-9 print / 978-1-7185-0477-6 ebook, $59.99). The book uses JavaScript/TypeScript (Node + Express) and Python (FastAPI) throughout, with all code available in a companion GitHub repository (github.com/jorshali/developers-guide-to-ai).
Structure - Five Parts
Part I gets a local model running via Ollama and streaming responses to a UI before the end of chapter 3 - no external API key required. Part II covers prompt engineering fundamentals and production patterns, focusing on prompts that behave consistently at scale rather than one-off tricks. Part III builds a complete RAG system: vector database setup with ChromaDB, chunking strategies, deterministic IDs, metadata preservation, and retrieval design. Part IV covers fine-tuning trade-offs, data preparation, and the decision curve between prompting, RAG, and custom models. Part V covers agentic workflows using LangGraph's ReAct pattern and extending agents with tool use - file I/O, API calls, real state persistence.
Authors and Technical Review
Jacob Orshalick (20+ years, independent consultant, co-author of Seam Framework), Jerry M. Reghunadh (20+ years, Senior Director at a global product organization, international conference speaker), Danny Thompson (host of The Programming Podcast, organizer of Commit Your Code conference). Technical review by Nikhil Kapoor (16+ years in AI/ML).
Positioning
The book explicitly targets developers who have used ChatGPT or Claude and want to build production systems, not toy demos. It covers provider-agnostic patterns first, acknowledging that specific models and APIs will shift while emphasizing engineering fundamentals - chunking strategy, retrieval design, fine-tuning cost curves, agent tool contracts - that remain load-bearing across provider generations. Available from No Starch Press, Amazon, and Barnes & Noble.
Key Points #
- 1The book covers LLMs, prompt engineering, RAG pipelines, fine-tuning, and autonomous agents in five sequenced parts, targeting JS/TS and Python developers with no prior ML background required.
- 2Authors emphasize durable engineering patterns over provider-specific tutorials, including when to choose RAG vs. fine-tuning vs. prompt engineering for production deployments.
- 3Published by No Starch Press (320 pages, ISBN 978-1-7185-0476-9) with a companion GitHub repo; co-authored by three practitioners with 50+ combined years across consulting, product engineering, and developer advocacy.
Scoring Rationale #
A well-structured practitioner book from a respected technical publisher covering production AI patterns (RAG, fine-tuning, agents). Relevant to the LDS developer audience but a niche publication event with limited broad market impact.
Practice interview problems based on real data
1,625 SQL & Python problems across 15 industry datasets — the exact type of data you work with.