Video Talk
Creating Machine Learning (ML) pipelines can often be complicated because of the tools and in-depth technical knowledge needed to create and deploy the right prediction models.
AutoML systems provide a black-box solution to these problems, looking for the right way to process and select features, choosing an algorithm, and fine-tuning the hyperparameters of the entire pipeline.
Being able to recall these autoML systems also from non-canonical model development environments, such as a DWH query shell, is possible, for a wide range of users, to harness the power of predictive analytics even without a deep machine learning experience.
