mohammad-agus/machine-failures-detection ? reverse-engineered prompt

Build me a simple machine failure detection app from this dataset that takes a few machine readings and tells me the chance of a failure.

I want the whole flow in one project, starting with data cleaning and training a few classification models, then picking the best one and saving it so it can be reused later. After that, make a small API that loads the saved model and returns a prediction from a /predict endpoint. It should be easy to run locally and also work in a container.

Then create a basic Streamlit page where I can enter the machine values in a form, send them to the API, and see the result clearly on the screen. Keep the interface simple and practical, like something you could use to test whether a machine is at risk. If you need to check current docs for Streamlit, Flask, or deployment details, go ahead and do that.