Amazon Web Services announced on October 22, 2026, the general availability of Amazon SageMaker AI Training Sessions, a serverless feature that lets developers create their own training loops. Supported approaches include supervised fine-tuning, reinforcement learning, custom loss functions, and multi-turn reinforcement learning, while AWS manages the GPUs, distributed execution, and infrastructure. The service is accessible from a local computer or SageMaker Studio through Python SDK clients and includes a cookbook with recipes for supervised fine-tuning and reinforcement learning. It supports open-weight models such as Llama, Gemma, Qwen, Nemotron, and gpt-oss, and is available in Northern Virginia, Oregon, and Mumbai. Pricing is based on the volume of tokens processed, with automatic encryption at rest in S3 and access governed by IAM managed policies.
Source: Read the original article
