Machine-Learning-Roadmap: a curated link list, not a runnable project
A roadmap for getting started with Machine Learning
At a glance
- What is it?
- Shanmukha Sainath's repository is a static GitHub Pages roadmap that points beginners at specific maths, ML and deep learning courses and books. It contains no code, no configuration and no tooling, so the decision to adopt it is a decision about a reading order, not about software.
- Who is it for?
- Adopt this roadmap if you are starting from zero and want one person's ordered list of maths, ML and deep learning material rather than a search engine result page. Do not adopt it if you need hands-on labs, deployment, MLOps or tabular data workflows; the README's topics list mentions tabular-data and computer-vision, but the body of the roadmap stops at courses and books.
- Can I use it commercially?
- Yes. MIT is a permissive licence: you can use, modify and sell software built on it, as long as you keep its copyright and licence notices.
- Is it still maintained?
- Yes. The repository last received commits 3 days ago.
- What is it written in?
- GitHub does not report a main language for this repository.
Answers come from the project's GitHub data, last synced on September 15, 2026, and from our analysis. They are not legal advice.
DEEP OPEN-SOURCE ANALYSIS
What this repository actually is
This is a Markdown README rendered as a website through GitHub Pages, with the homepage listed as shanmukh05.github.io/Machine-Learning-Roadmap. The author introduces himself in the README as an AI Engineer at KLA Corporation with a bachelors from the Electronics and Electrical Communication Engineering department at IIT Kharagpur. His stated reason for building it is stated plainly: there are numerous free and paid resources to learn Machine Learning, and having many options in hand confuses people. The repository exists to collapse that choice down to a short list. There is no source code in the material provided, no package manifest, no build step and no test suite. The primary language field is unknown, which is consistent with a documentation-only repository. If you are looking for a library to install, you are in the wrong place. If you are looking for a sequenced reading list maintained by one practitioner, this is the intended use.
The ordering is the product
The roadmap splits into Prerequisites, Machine Learning and Deep Learning, with a Frameworks and Libraries section and a What's Next section listed in the table of contents. Prerequisites are broken into Mathematics, Programming Fundamentals and Programming Language. Under Mathematics the README names Linear Algebra, Matrix Algebra, Probability and Statistics, and Calculus, and attaches one recommendation to each: MIT 18.06 for linear algebra, Khan Academy for matrices, Khan Academy for statistics and probability, and Khan Academy differential calculus for calculus. Programming Fundamentals gets a single entry, MIT 6.006 Introduction to Algorithms. Python is covered by a W3Schools tutorial. Every one of these is a link with a one-line justification. The value here is the sequence and the narrowing, not the depth of the annotations. The README asserts that MIT 18.06 is the best course for linear algebra basics and that the Khan Academy statistics course is the best available, which are opinions, not findings. They are presented as opinions by a named person, which is at least honest, but a reader should treat the word best as a preference rather than a comparison result.
Progress markers that are decorative, not functional
Between sections the README embeds images from geps.dev/progress with values 20 and 40. These render as progress bars in the GitHub view. They are hardcoded image URLs, not computed state. Nothing tracks what you have finished, nothing persists between visits, and the numbers do not correspond to any file in the repository. If you read the roadmap as a checklist, the bars will sit at 20 and 40 forever regardless of how much you have completed. This is a small thing, but it tells you what kind of artifact you are holding: a static document with visual flourishes, not an application. Anyone expecting a tracker, a quiz, or an interactive curriculum tree will be disappointed. The MIT licence covers the text and any assets in the repository, so you can fork it, reorder the sections, drop the progress images and publish your own version under the same terms provided you keep the licence notice.
Getting it running means cloning or reading
There is no install command in the supplied material. The realistic options are to read the rendered page at the homepage URL, or to clone the repository and open the README locally. Cloning follows the standard form for a GitHub repository: git clone https://github.com/shanmukh05/Machine-Learning-Roadmap.git, then open README.md in any Markdown viewer. Because the content is Markdown with raw HTML blocks, including ul, li and details elements, some renderers will show the collapsible Details sections differently from GitHub. If you want the same appearance as the published site, read it on GitHub Pages rather than in a local editor. There are no config keys to set, no environment variables, no dependency file and no build target. The only maintenance action available to a reader is a pull request against the README, which the author invites with the line Feedback and suggestions are welcome.
Where the roadmap stops short
The material covers foundations, classical machine learning and deep learning theory, and then lists Frameworks and Libraries and What's Next as sections without supplied content in the excerpt. There is nothing about data collection, labelling, feature stores, experiment tracking, model serving, monitoring or cost management. The repository topics include tabular-data, computer-vision, nlp, data-science and deep-learning, but the visible body of the roadmap does not map resources to those topics; the topics are discovery tags, not a curriculum structure. A second limitation is link rot. Several entries point to third-party PDFs hosted on personal or departmental file servers, for example the Bishop Pattern Recognition and Machine Learning PDF on a users.isr.ist.utl.pt address and the Geron book PDF on a WordPress uploads path. Those hosts can move or disappear without notice, and the repository has no link checker visible in the material. A third is currency: the README marks the Andrew Ng Machine Learning Specialization as a new course, and the deep learning list includes the 2020 NYU Deep Learning with PyTorch course by Yann LeCun, so parts of the list reflect a particular moment. There are no releases retrieved for this repository, which means there is no versioned snapshot to pin your reading to.
How it compares with a structured curriculum
The obvious alternative is a full curriculum with graded assignments, such as a university course sequence or a platform specialisation that bundles lectures, labs and a certificate. The difference in approach is that this roadmap curates and points outward, while a specialisation owns the material and sequences it with exercises and assessment. A learner on the roadmap has to assemble their own practice, because the README supplies links and one-line reasons, not problem sets. The trade-off runs both ways. A bundled specialisation gives you feedback and a defined finish line, but locks you to one vendor's teaching style and often to a paid tier. This roadmap lets you mix MIT 18.06 with the fast.ai course and the Goodfellow deep learning book at no cost, but it gives you no signal about whether you have actually learned anything. If you need external accountability, the roadmap alone will not provide it.
Maintenance, licence and what to verify
The repository is not archived and the last push recorded is 2026-09-07, so it is being touched. There are no releases, so there is no changelog to read and no way to tell what changed between visits except by diffing the README yourself. The licence is MIT, which permits reuse and modification with attribution and without warranty; that is a permissive arrangement for a document, and it means you can lift the structure into your own onboarding material. It does not mean the linked third-party resources are MIT licensed, and several of them are commercial courses on Coursera, Udemy and Amazon. Before you commit a study plan to this list, check three things: whether each linked course still exists at that URL, whether the free PDFs are still reachable from the hosts the README points to, and whether the Frameworks and Libraries and What's Next sections contain enough detail for your target role, since those are the parts most likely to be thin for someone heading into production work rather than research.
Editorial conclusion
Adopt this roadmap if you are starting from zero and want one person's ordered list of maths, ML and deep learning material rather than a search engine result page. Do not adopt it if you need hands-on labs, deployment, MLOps or tabular data workflows; the README's topics list mentions tabular-data and computer-vision, but the body of the roadmap stops at courses and books. Before relying on it, open the GitHub Pages site at shanmukh05.github.io/Machine-Learning-Roadmap and click every link, because several point to third-party PDFs and personal file hosts that the repository owner does not control.
Community notes