Build your own <insert-technology-here> (github.com)

This is a curated repository of hands‑on, step‑by‑step guides for rebuilding a huge range of technologies from first principles — everything from 3D renderers, game engines, web browsers, OS kernels and network stacks to BitTorrent clients, blockchains, containers and full-stack front‑end frameworks. It also collects machine learning and CV tutorials (neural networks from scratch, convolutional models for traffic‑sign classification, OCR, LSTM music generation, visual recognition with OpenCV), and language‑specific walkthroughs in C/C++, Python, JavaScript/TypeScript, Go, Rust, Java, C#, etc. Popular items include “Ray Tracing in One Weekend,” simple Redis/databases, tiny compilers/interpreters and container/VM reimplementations — all focused on learning by rebuilding. For AI/ML practitioners this is valuable as a compact lab of reproducible, low‑level learning projects: reimplementing backprop or a convolutional layer teaches numerical stability, gradient flow and memory/compute tradeoffs far better than high‑level APIs. Building data pipelines, Dockerized inference stacks, lightweight CV systems or custom model runtimes exposes real‑world issues (I/O, concurrency, quantization, profiling, deployment). The collection also supports experimentation — custom layer primitives, novel optimization tweaks, or building interpretable toy models — and doubles as a teaching curriculum for students and researchers wanting deeper intuition. (Be mindful: reimplementing protocols or cryptosystems carries security/ethical considerations.)