Amazon Adds Agentic Fine-Tuning to SageMaker AI Platform
Amazon Web Services has introduced agentic fine-tuning capabilities to its SageMaker AI platform, enabling enterprise developers to customize large language models through an automated AI agent workflow, according to The Decoder.
The new feature allows developers to fine-tune models including Meta’s Llama, Alibaba’s Qwen, DeepSeek, and Amazon’s Nova family of models using an automated AI agent that streamlines the model customization process.
The agentic approach automates configuration steps that previously required significant ML engineering expertise, allowing developers to tailor model behavior for enterprise applications without managing the full complexity of the fine-tuning pipeline, according to the report.
SageMaker operates in the enterprise AI platform market alongside Microsoft’s Azure Machine Learning and Google’s Vertex AI, where all three hyperscale cloud providers have expanded tooling for customizing and deploying large language models.
Amazon’s inclusion of models from multiple providers — including Alibaba’s Qwen and DeepSeek alongside its Nova models — reflects a multi-model approach on AWS that gives enterprise customers flexibility in choosing foundation models while remaining within the AWS ecosystem.
SageMaker is widely used by U.S. enterprises and developers for machine learning workloads.
