zenml

ZenML manages both classical ML pipelines and modern LLM-based agents in one framework, as emphasized in the README's key features for unified orchestration across the full MLOps lifecycle.

It runs pipelines on any backend like Kubernetes, SageMaker, or GCP Vertex via stacks, abstracting infrastructure complexity so users can focus on code, as described in the infrastructure abstraction feature.

The platform automatically containerizes code and tracks runs with metrics, logs, and metadata, ensuring reproducibility, which is a core part of ZenML's operationalization process.

ZenML integrates with existing ML tools such as MLflow, LangGraph, and Weights & Biases, allowing teams to orchestrate their preferred stack without replacement, per the tool integration section.