AI Engineering from Scratch
An open-source AI educational course project released under the MIT License, comprising 523 lessons organized into 20 stages. The curriculum emphasizes a stdlib-first philosophy and requires learners to implement every algorithm step by hand, covering the full AI engineering stack from mathematical foundations through model development to production deployment.
Timeline (last 90 days)
Each lesson of the AI Engineering from Scratch course produces a reusable artifact, such as a prompt, skill, agent, or MCP server.
The provider of the course materials admits that they have not fully studied the AI Engineering from Scratch course.
The entire course is divided into 20 phases, ranging from mathematical foundations to engineering implementation.
Version v2026.10 compiles all course content into six volumes of EPUB and PDF e-books, included with the release.
The site interface and course content now support eight languages: Chinese, Hindi, Spanish, Arabic, French, Portuguese, Turkish, and Vietnamese.
The course emphasizes a stdlib-first (standard library first) philosophy, requiring learners to implement each algorithm step by hand rather than calling ready-made library functions.
Learners can integrate the course into an AI coding assistant by executing the command npx skills add rohitg00/ai-engineering-from-scratch.
Via the /start-learning command, the system provides a placement quiz and a study plan.
Project resources include the site aiengineeringfromscratch.com and the GitHub repository rohitg00/ai-engineering-from-scratch.
The project now runs the tests included with each lesson through CI (continuous integration) and has fixed previously broken datasets, models, and links.
3 more timeline events
All Facts (13)
The design philosophy of the AI Engineering from Scratch course is to implement every important concept oneself before using any library.
70%UnverifiedThe provider of the course materials admits that they have not fully studied the AI Engineering from Scratch course.
50%UnverifiedEach lesson of the AI Engineering from Scratch course produces a reusable artifact, such as a prompt, skill, agent, or MCP server.
50%UnverifiedThe site interface and course content now support eight languages: Chinese, Hindi, Spanish, Arabic, French, Portuguese, Turkish, and Vietnamese.
50%UnverifiedThe course emphasizes a stdlib-first (standard library first) philosophy, requiring learners to implement each algorithm step by hand rather than calling ready-made library functions.
50%UnverifiedThe entire course is divided into 20 phases, ranging from mathematical foundations to engineering implementation.
50%UnverifiedLearners can integrate the course into an AI coding assistant by executing the command npx skills add rohitg00/ai-engineering-from-scratch.
50%UnverifiedVia the /start-learning command, the system provides a placement quiz and a study plan.
50%UnverifiedProject resources include the site aiengineeringfromscratch.com and the GitHub repository rohitg00/ai-engineering-from-scratch.
50%UnverifiedThe project now runs the tests included with each lesson through CI (continuous integration) and has fixed previously broken datasets, models, and links.
50%UnverifiedThe course provides implementations in four languages: Python, TypeScript, Rust, and Julia.
50%UnverifiedAI Story Builders 是 MIT 许可的完全开源项目，可在 GitHub 及 Microsoft Store 免费获取
50%UnverifiedVersion v2026.10 compiles all course content into six volumes of EPUB and PDF e-books, included with the release.
50%