This ensures the project of an MR's head pipeline is eager loaded, preventing an N+1 query problem from occurring when viewing the list of MRs of a project.
