MLflow

MLflow is the standard platform for the machine-learning lifecycle: experiment tracking (params, metrics, artifacts), a model registry with staging, and deployment tooling — now extended to GenAI with LLM tracing, prompt tracking, and evaluation. 28.4k stars, Apache-2.0, 19.9M PyPI downloads/month, under the Linux Foundation. Databricks shapes the roadmap and the UI shows its age, but nothing else covers the full lifecycle with this ecosystem. The default 'serious ML team' choice.

```bash

pip install mlflow

mlflow ui # dashboard at :5000

```

```python

import mlflow

with mlflow.start_run():

mlflow.log_param("lr", 0.01)

mlflow.log_metric("acc", 0.95)

```

Docs: mlflow.org/docs.