Streamlining generative AI development with MLflow v3.10 on Amazon SageMaker AI

This article announces MLflow v3.10 support in Amazon SageMaker AI MLflow Apps, introducing enhanced capabilities for generative AI development and experiment tracking.

• MLflow 3.10 improves tracing for multi-turn workflows and LLM framework integration
• New mlflow.genai.evaluation() API provides built-in metrics for relevance, faithfulness, correctness, safety
• Enhanced observability with granular trace filtering, richer metadata, pre-built performance dashboards
• MLflow workspaces organize artifacts across teams and projects at scale
• Easy setup via SageMaker Studio console, AWS CLI, or API with default MLflow 3.10 provisioning
• Existing MLflow servers can migrate to SageMaker AI using serverless MLflow

MLflow v3.10 on SageMaker AI provides enterprise-grade infrastructure for tracking experiments, monitoring generative AI performance, and maintaining governance across AI applications.