Library / SDK
Observability & EvaluationFine-tuning & Training
MLflow
1Manage 2the 3end-to-end 4machine-learning 5experiment 6lifecycle
the six Ws · specification
W1
Who
Maintained by the MLflow project (originally Databricks), under the Linux Foundation.
W2
What
A platform for experiment tracking, model packaging, a registry, and LLM tracing and evaluation.
W3
Where
An open-source library and self-hostable tracking server; also managed services.
W4
When
Reach for it when tracking experiments, versioning models, and instrumenting GenAI apps.
W5
Why
The de facto open standard for ML lifecycle management with a large ecosystem.
W6
With
Integrates with scikit-learn, PyTorch, TensorFlow, LangChain, and OpenAI.
W7
Watch
Community-governed (Apache-2.0), vendor-neutral, and fully self-hostable. Data stays wherever you host the server.