100-days-mlops-kodekloud

Create and Organize MLflow Experiments

Problem

The xFusionCorp Industries ML platform team is onboarding two new ML projects and needs each one organised under its own MLflow experiment rather than sharing the Default experiment. Your task is to register both experiments through the MLflow UI and tag them with the owning team.

1. The MLflow tracking server is already running on port 5000. The MLflow UI button at the top of the lab can be opened to view the dashboard. One seeded experiment (legacy-models) is listed alongside the platform-created Default—both act as reference material and must not be modified.

2. Using the MLflow UI, register two new experiments with the experiment-level metadata below. The task is complete when both records satisfy every bullet.

   • fraud-detection

     • Experiment-level description is a non-empty string describing the project (any phrasing).
     • Experiment-level tag: key team, value ml-platform.

   • churn-prediction

     • Experiment-level tag: key team, value analytics.

The result can be confirmed in the MLflow UI: both new experiments appear in the left-hand list, with the description and tags visible on each experiment’s page.

Solution

Pretty simple. Just click on UI button and follow the instructions.

• You have to create two new experiment fraud-detection and churn-prediction with descripting(anything you can put) and given tags with key-value.

This site is open source. Improve this page.