The scope of the work presented in this paper was to apply the latest open source CFD achievements to design a state of art, direct-injection (DI), heavy-duty, natural gas-fueled engine. Within this context, an initial steady-state analysis of the in-cylinder flow was performed by simulating three different intake ducts geometries, each one with seven different valve lift values, chosen according to an estabilished methodology proposed by AVL. The discharge coefficient (Cd) and the Tumble Ratio (TR) were calculated in each case, and an optimal intake ports geometry configuration was assessed in terms of a compromise between the desired intensity of tumble in the chamber and the satisfaction of an adequate value of Cd. Subsequently, full-cycle, cold-flow simulations were performed for three different engine operating points, in order to evaluate the in-cylinder development of TR and turbulent kinetic energy (TKE) under transient conditions. The latest achievements in open source mesh generation and motions were applied, along with time-varying and case-fitted inizialization values for the fields of intake pressure and temperature. Finally, the supersonic injection of natural gas was coupled with full-cycle computations, both to evaluate influence of the injected fuel over the in-cylinder charge motion and to predict air-fuel mixing homogeneity in terms of equivalence ratio and homogeneity index (HI). Three specific engine operating points were simulated and different combinations of turbochargers and valve lift laws were tested. Validation was carried out by comparing computed data with the results achieved by commercial CFD simulations deployed under the same conditions. However, for a proper assessment of the computed parameters a detailed set of experimental data is required.