Substack, a popular online publishing platform, uses machine learning (ML) for various purposes such as image generation, content categorization, content recommendation, semantic search, and audio transcription. To manage these ML workloads efficiently, Substack has moved its inference tasks to the Substrate platform. Previously, they had tried using other tools but found them slow and expensive. Substrate offers a polished API for performant inference across multiple models, enabling Substack to create large-scale, modular, multi-inference AI workflows with greater speed and flexibility.