# SDML bookclub: years of machine learning meetup notes kept in one repository > The SanDiegoMachineLearning bookclub repository collects the notes, slides, schedules and recording links from a San Diego machine learning group's reading sessions, one markdown document per book, going back to 2021 and forward to a course series that began in March 2026. It is a reading archive rather than a library, with no licence declared and most of the value living in links to videos and slides hosted elsewhere. **SanDiegoMachineLearning/bookclub** — Notes and links from the book club meetings - Repository: https://github.com/SanDiegoMachineLearning/bookclub - Stars: 804 · Forks: 150 - Language: Jupyter Notebook - License: not declared - Published: 2026-10-20 · Updated: 2026-10-20 · Language: en - Canonical page: https://hysenlabs.com/projects/sandiegomachinelearning-bookclub ## A reading archive, not a code library The repository description is one line: notes and links from the book club meetings. The README opens by saying the same, notes and links from the SDML book club meetings, and then adds a pointer to a separate SDML talks repository for slides and videos from other, non-book machine learning talks. That distinction is the fastest way to understand the project. This is the San Diego Machine Learning group's archive of its book reading sessions. The primary language is recorded as Jupyter Notebook, which tells you that some of the content is executable rather than purely prose, but the spine of the repository is a set of markdown documents, one per book. Each book gets a top level section in the README with a cover image and a link to its own document. For AI Engineering, the README points at ai-engineering.md and says the notes, slides and videos from the meetup sessions for that book can be found there. For LLM Inference Illustrated it points at llm-inference.md and, with a detail worth noting, describes that file as holding links to video recordings rather than notes. Computer Vision, Mathematics for Machine Learning and Understanding Deep Learning are described more fully, with the schedule, notes, slides, record links and extra resources all collected in the same document. So the convention is consistent: a per-book markdown file that gathers the schedule, the notes, the links to recordings and any extra files. The repository is the index, and the per-book documents are the content. For a newcomer, this is a useful template. For an engineer looking for working code, it is worth being clear that the interesting artefact is often a link out to a video or a slide deck rather than a notebook in the tree. ## Supporting directories beside the markdown Several books come with a directory of their own, and those directories carry the files that do not fit in a markdown file. The top level listing shows homl/, cv/, graph/, mml/, tysql/ and udl/ sitting alongside the markdown documents, plus an images/ directory holding the cover art the README renders. The directory names map onto the books: homl for Hands-On Machine Learning with Scikit-Learn, Keras, and TensorFlow, cv for Computer Vision, graph for Machine Learning with Graphs, mml for Mathematics for Machine Learning, tysql for the Learning SQL series, and udl for Understanding Deep Learning. There is also a standalone PDF in the root, Exploring the Reinforcement Learning Techniques in AlphaGo.pdf, which is a paper rather than a book and does not belong to any of the reading groups. The split between prose and directory content is a practical one. Markdown files are diffable, reviewable in a pull request and readable on GitHub without any tooling. A directory of notebooks or images is not, and for a group that accepts outside contributions, the pull request workflow is what keeps the archive trustworthy. One entry in the root listing is a document for a book that the README no longer lists in the visible portion, deep-learning-with-pytorch.md, which suggests the archive also keeps earlier work around rather than pruning it when a session ends. ## The session schedule tells you what the group was reading The README records when each series started, and those dates are the clearest signal of how the group works over time. LLM Inference Illustrated started in March 2026, which makes it the most recent entry, and the Hands-On Machine Learning third edition series started in August 2025. Probabilistic Machine Learning: Advanced Topics started in January 2023. The Learning SQL series, using Sams Teach Yourself SQL in 10 Minutes, started in September 2022. Quantum Computing, using Dancing with Qubits, started in August 2022, and Machine Learning with Graphs started in January 2022. The second edition of Hands-On Machine Learning started in October 2021, which is the earliest start date in the visible list. Read as a sequence, that is roughly five years of continuous activity, one book or course at a time, with the group moving from classical topics through probabilistic methods and SQL into the large language model literature. The dates also show the pace: a book takes months rather than weeks, and the group returns to the same author more than once, since Hands-On Machine Learning appears twice, three years apart, in different editions. That last detail is the kind of thing a reader can reuse. If you are running a group, revisiting a book when a new edition appears is a cheap way to keep a cohort engaged, because the new edition gives the group a reason to meet again. The Generative AI with Large Language Models entry does not name a start date, which suggests it is handled slightly differently from the book groups. ## Course slides and video hosted outside the repository The Generative AI with Large Language Models entry documents a course rather than a book, and it is the one section in the README that lists per-week links inline. The group is going through the course by DeepLearning.AI on Coursera, and the README enumerates the kickoff meeting and the first three weeks, each with a slides link to a Google Slides document and a video link to a YouTube recording. The links are given as full URLs for each week, with the kickoff, week one, week two and week three each pointing at their own slides deck and recording. The Algorithms to Live By entry follows the same rule from the other direction. Instead of pointing into this repository, it says the slides are in a separate talks repository and the video is on YouTube. The README already told you at the top that non-book talks live in the SDML talks repo, and this entry links to an individual person's talks repository for the slides. Both patterns point at the same underlying constraint. Video recordings and slide decks are large, and Google Slides and YouTube are where the group already had them. The repository keeps the index and the notes, and borrows the rest. For someone reusing this content, the practical implication is that a link in these documents is a promise, not a copy. Sessions recorded on platforms with their own retention and editing policies can disappear, and a document with a link to a video is only as durable as that link. ## What you can and cannot take from a notes repository There are real limits to this kind of archive, and they are worth being explicit about. The first is coverage. These are meeting notes. A group that meets for a couple of hours a week covers a chapter or two, and the notes reflect that. They are a guided reading companion with the particular insights of the people in the room, not a substitute for the book. Where a book teaches a method properly, a meetup will usually summarise it and move on. The second is uneven depth. The README's own descriptions differ. One entry promises notes, slides and videos, another only links to video recordings, a third mentions a schedule plus notes plus record links plus extra resources. That inconsistency is natural in a volunteer archive and it means you should open the document before planning a session from it. The third is the licensing position. The repository lists no licence, which means the default copyright rules apply and there is no permission statement granting reuse. The books themselves are copyrighted works, and the notes are the group's own contribution on top of them. Any plan to publish a derivative work from these documents, or to reuse the notebooks in a course, needs a decision that the repository does not make for you. That is not legal advice, it is the reason to check before you copy. The fourth is that the content is anchored to a specific group's pace and to specific books that were current at the time. A 2022 session on a framework or a course has an expiry date built into it. ## Maintenance, activity and the shape of the archive The repository is not archived, and the last push was on 2026-10-13, which is consistent with a group that is still meeting. There are no GitHub releases, which is the normal arrangement for a collection of documents rather than a package: there is nothing to version as an artefact, and the history in git is the version history. The default branch is master, and the repository is the SanDiegoMachineLearning organisation's own, with a README in English. The organisation splits its output across at least two repositories, this one for book reading sessions and the talks repo for other machine learning talks, and individual contributors sometimes keep their slides in their own talks repositories, as with the Algorithms to Live By entry pointing at a separate account. That scattering is the main structural weakness. A group that keeps its book notes in one place but its slides and videos across personal accounts and third-party platforms is a group with an index and a set of links. For an archive that intends to outlive the people who made it, the notebooks and the markdown in this repository are the durable part, and everything else is a rental. What is durable here is still worth something. Five years of consistent, well-organised notes on a coherent curriculum, with a schedule for each series and a cover image for each book, is more than most reading groups produce, and the per-book markdown convention is a small piece of engineering that a new group could copy in an afternoon. ## Conclusion Use this repository if you are running a reading group or teaching from books and want a working example of how a group documents its sessions, or if you want the community notes for one of the books it covers before buying it. Do not treat it as a course: there is no licence declared, the content is other people's copyrighted work summarised at meetup pace, and the slides and videos live on Google Slides, YouTube and Coursera rather than in the repository. Before reusing any of it, check the file at the top of the repository you want to borrow from and confirm what the group itself intended, and read the per-book document in full, since the schedule and the discussion notes for different sessions are uneven in depth. ## FAQ ### What is in the SanDiegoMachineLearning bookclub repository? Notes and links from the SDML book club meetings, organised as one markdown document per book. Each document gathers the schedule, notes, links to recordings and extra resources for the sessions on that title. ### Which books has the SDML bookclub covered? The README lists AI Engineering, LLM Inference Illustrated, Hands-On Machine Learning with Scikit-Learn, Keras and TensorFlow in its third and second editions, Computer Vision, Mathematics for Machine Learning, Understanding Deep Learning, Probabilistic Machine Learning: Advanced Topics, Learning SQL, Quantum Computing, Machine Learning with Graphs and Reinforcement Learning. ### Where are the slides and videos for the bookclub sessions? Mostly outside the repository. The README says slides and videos from other, non-book machine learning talks live in the separate SDML talks repo, and the Generative AI with LLMs course entry links to Google Slides decks and YouTube recordings for each week. ### Can I reuse the notes and notebooks from the bookclub repository? The repository declares no licence, so the default copyright rules apply and no reuse permission is granted in the repository itself. The books are copyrighted works in any case, so check with the maintainers and the rights holders before republishing anything. ### How often does the SDML bookclub add new sessions? The repository is not archived and the last push was on 2026-10-13. Series start dates in the README range from October 2021 for the second edition of Hands-On Machine Learning to March 2026 for LLM Inference Illustrated, so sessions run continuously rather than in bursts. ## Sources - [Issues](https://github.com/SanDiegoMachineLearning/bookclub/issues) - [README](https://github.com/SanDiegoMachineLearning/bookclub/blob/master/README.md) - [SanDiegoMachineLearning/bookclub on GitHub](https://github.com/SanDiegoMachineLearning/bookclub) --- Hysen Labs editorial analysis, written from the project's own repository and release notes. Cite the canonical page: https://hysenlabs.com/projects/sandiegomachinelearning-bookclub