We consider the optimization of container loading at three intermodal logistics platforms (ILP) of a large European car manufacturer (ECM). The decisions focus both on the loading day of each container and on its flling with the products in inventory, which are gradually received over the week from inland suppliers. The objective is either to reduce the largest inventory level needed in the ILP, or to smooth the weekly workload. We develop a solution methodology that allows the handling of complex loading constraints related to dimensions and weight of the products. We model the problem as a mixed integer linear program and we develop a decomposition heuristic to solve it. We perform extensive computation tests on real instances provided by ECM. Compared with current industrial practices, our solutions yield an average improvement of 46.8% for the inventory reduction and of 25.8% for the smoothing of the workload. Our results highlight the benefit of jointly optimizing container loading and operations scheduling.