SageMaker AI integrated serverless customization for NVIDIA's open-weight Nemotron 3.5 Lightning model, which has a hybrid Mixture-of-Experts architecture with billions of parameters. This allows users to tailor the foundation model for their specific data and applications using supervised, preference, or reinforcement fine-tuning methods, optimizing accuracy, alignment, and task performance.
By managing infrastructure and training automatically, SageMaker reduces operational overhead and cost. The service is available in multiple global regions and supports model adaptation workflows via SageMaker Studio or Python SDK, enabling efficient deployment of customized AI models without managing dedicated compute clusters.
Source & further reading
Amazon SageMaker AI now supports serverless model customization for NVIDIA Nemotron 3.5 Lightning ↗Source published 2026-09-16. This report was prepared with AI assistance from the linked primary-source material. Claims and limitations should be checked against the original announcement or research.