sejaljain-27/ml-assistant ? reverse-engineered prompt

Build me an AI powered ML assistant web app where a user can upload a CSV file and instantly get a clear analysis of the dataset.

It should tell me whether the problem looks like classification or regression, check for missing values, duplicates, outliers, high cardinality columns, imbalance, and correlated features, then give the dataset a simple health score with warnings and helpful recommendations. I want a polished dashboard that shows charts for things like class distribution, correlation, missing values, feature importance, and model comparison, plus a step by step pipeline view so it is easy to see what passed or needs attention.

The app should also train a few baseline models for the detected task, compare them with the right metrics, and suggest which models and preprocessing steps make sense. Keep the design clean and premium looking, and make the frontend talk to a backend that handles the analysis. If you need current docs for any libraries, look them up online.