MLflow is an open source platform to manage the ML lifecycle, including experimentation, reproducibility, deployment, and a central model registry. MLflow currently offers four components: MLflow Tracking, MLflow Projects, MLflow Models, Model Registry
This is nice, however lots to be desired : how do you incorporate feature stores? Also they need to decide where this fits into the ML lifecycle. Is this a research or production phase thing? I like that you can easily run scripts from notebooks.
I wish there was a tool that combined Kedro and mlflow. Kedro pipelines are very good, with modularity, visualization and data catalog with version control and I/O abstraction, combined with mlflow tracking, model and model registry is all you need. Kedro and mlflow will complete each other.
Hello Gibrail! These office hours will be hosted biweekly on Thursdays. You can join our community to be notified about future MLflow events here: community.linuxfoundation.org/mlflow/