Detailed simulation of showers in calorimeters is often the most time-consuming part of computing for high energy physics (HEP) experiments. Instead of the expensive multi-step particle tracking computation, one can develop models that generate energy deposited in the calorimeters according to some parameterisations. Machine learning (ML) techniques are employed to reproduce particle showers. A prototype of the example application for fast shower simulation has been developed and integrated within the standard toolkit for HEP simulation, Geant4. It demonstrates how to produce data for training of the ML model, how to train the model, and use it in the simulation alongside the detailed simulation.