Timber

Timber is doing something I didn't know I needed until I saw it: compiling XGBoost, LightGBM, and scikit-learn models into lean C99 inference code with a single command. Classical ML models are the actual backbone of most production systems — but they get zero deployment love compared to neural networks, and that's genuinely unhinged. This is the move that gives tree-based models the same first-class treatment LLMs got with Ollama, and if you're still shipping Python at inference time, I have questions.