An additional lesson to note is how these competitions test more your ML Model Deployment skills rather than your ability to invent a new model in less than a week. Therefore, it’s worthwhile to learn about how to serve your models on the cloud on Azure or AWS, and how to productionize your ML data pipeline (aka you give me the raw input data, and I have setup several python functions to transform it, send it to the model, and output you the model’s results that I can include in the backend of a large scale web or mobile app)
Im doing one for the first time soon so thanks for the great general advice.
An additional lesson to note is how these competitions test more your ML Model Deployment skills rather than your ability to invent a new model in less than a week. Therefore, it’s worthwhile to learn about how to serve your models on the cloud on Azure or AWS, and how to productionize your ML data pipeline (aka you give me the raw input data, and I have setup several python functions to transform it, send it to the model, and output you the model’s results that I can include in the backend of a large scale web or mobile app)