All of the Concrete ML models are compatible with the scikit-learn Estimator API. Thus you can use our models in a pipeline, just as you would use the models from scikit-learn. Note that an additional parameter is added to some models:
n_bits. While you can leave this to the default value, you can also change this in the pipeline or in GridSearchCV if you want to experiment.
Check out the grid search example from the docs !