Hey everyone, I'm starting a project called CatAI—a native AI model engine built entirely from scratch in C++. I want to skip the bloated standard frameworks and write the core tensor math and memory layouts directly. Because doing a raw scratch build is complex systems work, I'm looking for other C++ devs or math nerds who want to collaborate, brain-dump architecture ideas, or help guide the backend logic. If you love bare-metal optimization, low-level memory management, or custom neur...