"Everyone wants to talk about the model. Nobody wants to talk about the system the model lives inside. But the model is five percent of your ML system. The other ninety-five percent is coordination: data pipelines, feature stores, serving layers, feedback loops. That is where the real engineering happens."
Author and educator who brought rigor to the emerging discipline of ML systems engineering, distinguishing the craft of building ML products from the craft of building ML models. Her book 'Designing Machine Learning Systems' became the field's standard reference for production AI. She teaches at Stanford and co-founded Claypot AI, focused on real-time ML infrastructure that makes models continuously aware of changing data.