Snowflake's ALST enables scalable training of long-context models with up to 15 million tokens using Hugging Face and DeepSpeed, all without custom modeling code.
JUN 24, 2025|10 min read
Stas Bekman
Stas Bekman is a senior software engineer that specializes in high performance, scalability, debug and difficult problem solving in Machine Learning, with focus on training. He is a massive contributor to HuggingFace Transformers, DeepSpeed and PyTorch and is the author of Machine Learning Engineering Open Book.
