The best reference material is the kind you keep reaching for long after you have finished the book.
Sage, Reference-Hoarding AI Agent
Part Overview
Four appendices of reference material. Appendix A: Mathematical Foundations covers the math prerequisites every transformer relies on, linear algebra, probability, calculus, and information theory. Appendix B: Course Syllabi contains five tested course tracks with week-by-week schedules and assessment rubrics. Appendix C: Reading Pathways contains per-audience reading guides for engineers, researchers, founders/PMs, and self-study learners. Appendix D: Agents That Helped to Write This Book is the roster of the 42 specialist AI agents in the writing pipeline that produced this manuscript, with a card per agent explaining what each one owns. Framework, tooling, and ops material that previously lived in appendices has been consolidated into the part-specific Tools of the Trade chapters so it sits next to the content that uses it.
These appendices serve two purposes: they provide prerequisite refreshers you can consult before diving into specific chapters, and they offer practical framework guides you will return to when building real projects. Think of them as the book's toolbox.
The essential linear algebra, probability, calculus, and information theory that power every transformer. Six sections covering vectors and matrices, probability distributions, derivatives and gradient descent, entropy and KL divergence, and information theory for language models.
Open Appendix A →Five tested course tracks (undergraduate engineering, undergraduate research, graduate engineering, graduate research, professional bootcamp) with week-by-week schedules, prerequisites, and assessment rubrics for instructors building courses on this book.
Open Appendix B (Course Syllabi) →Per-audience reading guides for engineers, researchers, founders/PMs, and self-study learners. Each pathway tells you what to read in what order, roughly how long it takes, and what you can do after.
Open Appendix C (Reading Pathways) →The roster of 42 specialist AI agents that produced this book. Each card has an avatar, role tag, focused bio, and the single core question the agent asks. From Chapter Lead and Curriculum Alignment through Publication QA and Hands-On Lab Designer.
Open Appendix D (Agent Roster) →A standalone mini-book on PyTorch in ten sections: tensors, autograd internals, nn.Module patterns, data pipelines, the training-loop deep dive, mixed precision, distributed training with FSDP, torch.compile and the profiler, debugging recipes, and saving, loading, and deployment.
Open Appendix E (PyTorch Reference) →The compact signal-processing toolkit needed to read Chapter 20: sampling, framing, and windows; the DFT and FFT; the mel scale, log-mel spectrograms, and MFCC; and a Z-transform primer. The math prerequisites for every audio model in the book.
Open Appendix G (Signal Processing for Audio) →