This talk explores the unique challenges of productionizing foundation models compared to traditional machine learning models. Despite sharing some core principles, foundation models introduce new complexities due to their open-ended nature, advanced capabilities, and computational demands. Key changes include shifting from closed-ended to open-ended evaluation, from feature engineering to context construction, and from structured data to unstructured data.
Chip Huyen works to accelerate data analytics on GPUs at Voltron Data. Previously, she was with Snorkel AI and NVIDIA, founded an AI infrastructure startup (acquired), and taught Machine Learning Systems Design at Stanford. She’s the author of the book Designing Machine Learning Systems, an Amazon bestseller in AI.
