A free, MIT-licensed course of 542 lessons on AI engineering, implementing model internals, retrieval, and agent runtimes from scratch.
Implement a ChatGPT-like LLM in PyTorch from scratch, step by step
Implementing LLM architectures from scratch in PyTorch
A straightforward method for training your LLM, from downloading data to generating text.
A 2.78-trillion-parameter Kimi K3 running inference on a single CPU in 8.24 GB of RAM. Portable C99: no BLAS, no framework, no GPU.
A survivors roguelite running on a from-scratch World of Warcraft 1.12.1 client engine (Rust + Bevy).
Implement a reasoning LLM in PyTorch from scratch, step by step
Agent工程师最全学习路径 · 从零精通 AI 工程 · 20 阶段 542 课 · 中文全量翻译 + 配套站点 + 动画讲解视频 · 如何成为 AI Agent 工程师的修成指南
LeetCode for PyTorch — 65 ML/AI interview problems from real interviews at Google, Meta, Anthropic. Jupyter notebooks, an auto-grader, and an MCP AI tutor.
Don't memorize ML. Understand it from first principles including Inference, Cuda, Deep Learning, Data Science and Reinforcement Learning
Learn harness engineering by building Claude Code from scratch. Free, open-source course: 8 articles, 6 videos, 1 codebase.
A from-scratch, standalone implementation of the Zygisk Next module
A from-scratch Windows 98 display driver with hardware Direct3D OpenGL & Glide on the Intel GMA 950 and S3 ViRGE/DX and Trio3D/2X, 2D on S3 Trio32/64, and a VESA fallback.
End-to-end project training and building AI models / LLMs to detect AI-written text
Vantage: a from-scratch LuCI theme and live dashboard for OpenWrt 25.12
600 行 TypeScript 写成的超级迷你版 pi,让你轻松从 0 写出属于你的 pi-agent
From teacher to tiles — a from-scratch LLM distillation & serving engine: custom Triton/CUDA kernels, FSDP distillation, paged-KV continuous batching, speculative decoding, a Rust gateway, a JAX oracle, and interpretability tooling.
Red Alert 2 + Yuri's Revenge running natively on iPhone and iPad. EA never released RA2's source, so this builds on Chrono Divide's from-scratch TypeScript engine — proprietary, used with the author's permission, non-commercial only — wrapped in a Swift/WKWebView shell. Bring your own retail copy; no game assets included.
A skill to kill from-scratch coding — Claude checks real arXiv prior art before it designs a new architecture.
A 25.7M-parameter GLM-5.3-Flash-inspired model trained from scratch and improved with executable-reward RL.
PyTorch 기반 한국어 LLM 구현 및 학습 프로젝트
Wire v9: from-scratch ultra-low-latency HFT stack in Rust and C. 24 crates: TCP/IP, AF_XDP, VFIO PCIe DMA, LMAX Disruptor, AVX2/AVX-512 SIMD, flat-array matching engine (11 cyc), HugePage slab (1.5 ns), lock-free MPMC, CME MDP 3.0 SBE, eBPF XDP risk. Zero heap allocs on the hot path. 46/46 tests pass.
PSX recompilation of Bloody Roar II (SLES-01722 EU / SCUS-94424 USA). Anti-copyright: bring your own disc. Dual-region builds, from-scratch widescreen + curated mods.
Independent desktop operating system with its own kernel, userspace and native BoringWM, written primarily in C
A single executed notebook building and training a small GPT with RoPE and a phase-aware sparse MoE experiment
A from-scratch reimplementation of the Knight Online game server and client
A from-scratch C++ implementation of Ray Tracing in One Weekend, extended with CUDA GPU acceleration, physically based rendering, progressive sampling, and procedural scenes.
A fast, memory-efficient variant effect predictor written in Rust: a from-scratch reimplementation of the Ensembl Variant Effect Predictor (VEP) that takes the same inputs and flags, emits VEP's consequence vocabulary in VEP's output formats, and annotates population-scale callsets two orders of magnitude faster than the Perl implementation.
Independent from-scratch implementation of the Kimi K3 architecture (arXiv:2607.24653v1): KDA, NoPE, latent-space MoE, and the systems co-designs. Table 1 reproduced to 0.09%.
Neural network from scratch in pure Python + NumPy that recognises handwritten digits (MNIST), with an educational GUI: watch it learn, tweak it while it trains, look inside every layer.
GLM-5.3 (full) on 3-6 DGX Sparks (GB10): a from-scratch CUDA inference engine
Build a World Model from Scratch — the Vizuara AI lecture series: slides, from-scratch code, and Colab notebooks
从零手写一个 AI Agent 的教学型 Python 项目:不依赖任何智能体框架实现核心循环,用约 3700 行可通读的代码,把 Agent 拆成乐高积木式的基础部件——模型调用、工具协议、执行上下文、记忆、RAG、Skills、E2B 沙箱与多智能体编排,逐层拼装、每一步都能追到源码。附 20 个可运行示例、21 章系统课程与 30 篇分层文档,配有单元与集成测试、mypy 严格类型检查。学操作系统要写迷你 OS,学数据库要写迷你数据库,学 Agent 就该亲手写一个:先建知识底座,再上生产框架。适合初学者、大学生与用过框架想补原理的工程师。MIT 开源,Python 3.13。
Machine learning from scratch with Python and NumPy. Explore 12 chapters covering regression, neural networks, trees, clustering, SVMs, and autodiff with mathematical derivations, interactive notebooks, gradient verification, and 59 automated tests. No scikit-learn or PyTorch.
A from-scratch homebrew SDK for Xbox 360 titles on the retail kernel, with a GPU layer that drives Xenos directly
Receive files from Android Quick Share on a jailbroken Kindle - no cable, no router, nothing on the phone. Includes a SoftAP mode and a from-scratch Go implementation of the Nearby Share protocol.
Local LLMs that run lean, load instantly and keep getting smarter. A from-scratch runtime + model format in Zig: zero-copy mmap, CPU/CUDA/Vulkan engines, OpenAI-compatible server, 2:4 sparsity, Net2Net growth, in-file version history, self-learning loops.
Native, notarized Android File Transfer for macOS and Linux — a from-scratch Swift MTP engine. No Electron, no drivers, no cloud.
Build an AI Coding Agent from Scratch in Pure Python, Step by Step | 用纯 Python 从零搭建 AI 编程 Agent,边实现边理解其工作原理