Multilevel Modeling: How to predict the parmeters of the lower model in the upper model

I am new to Stan and I am trying to fit a multi-level model.
The lower model should predict log (y + 0.1). The upper model should predict the coefficients of the lower model using the additional features (feature1 and feature2).

Does anyone have an idea on how to do this?