Open-source AI Engineering from Scratch curriculum builds every algorithm from raw math, agent-tutored in your terminal
ghumare64 · x · 2026-10-06
Engineer Rohit Ghumare maintains the open-source (MIT) curriculum "AI Engineering from Scratch," a full path from linear algebra to autonomous agent swarms.
- Philosophy: every algorithm—backprop, tokenizer, attention, agent loop—is derived from raw math and hand-implemented before any framework appears; no fragmented tutorials, no copy-paste deploys.
- Format: each lesson follows one loop (read problem, derive math, write code, run tests, keep the artifact), in four languages (Python/TypeScript/Rust/Julia) across four learning paths.
- Workflow: learn in your terminal via npx skills add rohitg00/ai-engineering-from-scratch; SKILL.md-capable agents like Claude, Cursor, and Codex act as tutors with placement quizzes and interactive lessons. Free and runs on your own laptop.
More from coding & agent
- OpenCode hits 17M MAUs just 16 months after launch — ycombinator · 2026-10-06
- Claude Code 2.1.291 Fixes Session-Drop Regressions, System Prompt Tokens Jump 15% — ClaudeCodeLog · 2026-10-06
- Claude Code 2.1.291 fixes permission-prompt and message-loss regressions — ClaudeCodeLog · 2026-10-06
- Claude Code 2.1.291 fixes cloud session and message-loss regressions — ClaudeCodeLog · 2026-10-06
- AI agent architecture explained: the 7 modules from perception to observability — ZabihullahAtal · 2026-10-06
- Memoria 1.0: a local, model-agnostic LLM memory layer hitting 89.8% Recall@1 on LongMemEval — kitkatz69 · 2026-10-06