AI & TinyML

This subdomain collects TAMCPP's AI / TinyML engineering labs. Like Embedded Development and Networking, you won't find concept quick-reference cards here — only hands-on projects of the "build a working thing from scratch" kind. The goal is to express neural-network inference in modern C++ in a way that's clear, controllable, and explainable, and along the way to weld std::array / std::span / constexpr / template dimension constraints / exception-free error handling / no heap allocation / static weights into one coherent piece.

Projects

• TinyInferCpp-Lab — Build a toy neural-network inference engine from scratch in C++23 (Input → Dense → ReLU → Dense → Argmax): float32, fixed structure, and a core path with no heap allocation, no exceptions, and no RTTI. PC closed loop first, with STM32 deployment as a follow-up stage.