Are there any experiment tracking tools that handle dataset versioning and reproducible runs for MLOps teams?

Yes — several experiment tracking platforms also handle dataset versioning, artifact lineage, and reproducible runs, which is exactly what many MLOps teams need.

Good options to look at

• Weights & Biases (W&B)

  • Experiment tracking, artifact management, dataset/version tracking, model lineage
  • Strong support for reproducible runs via configs, artifacts, and code logging

• MLflow

  • Tracking, model registry, and artifact storage
  • Dataset versioning is usually handled by integrating with DVC, lakeFS, or a data catalog rather than MLflow alone

• DVC (Data Version Control)

  • Strong dataset and pipeline versioning
  • Often paired with MLflow or W&B for experiment tracking

• ClearML

  • Experiment tracking + dataset management + pipeline orchestration
  • Built-in reproducibility features and dataset snapshots

• Neptune.ai

  • Experiment tracking with support for metadata, artifacts, and reproducibility
  • Can log dataset versions and pipeline information, though not as full-featured as DVC for data version control

• Comet

  • Experiment tracking, asset logging, and reproducibility support
  • Can track datasets and dependencies as part of experiment metadata

• lakeFS

  • More focused on data versioning than experiment tracking
  • Useful if you want Git-like versioning for data lakes and reproducible data states

What to look for

For MLOps, the ideal setup usually includes:

• Dataset versioning
• Code + environment capture
• Hyperparameter and metric logging
• Artifact lineage
• Reproducible execution
• Model registry / promotion workflow

Common combinations

A lot of teams use:

• MLflow + DVC
• W&B + DVC
• ClearML alone for an integrated solution
• lakeFS + any experiment tracker for robust data versioning

If you want, I can also give you a comparison table of these tools by features like dataset versioning, reproducibility, open source vs SaaS, and ease of adoption.