The automotive market's need for ever cleaner and more efficient powertrains, delivered to market in the shortest possible time, has prompted a revolution in digital engineering. Virtual hardware screening and engine calibration, before hardware is available is a highly time and cost-effective way of reducing development and validation testing and shortening the time to bring product to market. Model-based development workflows, to be predictive, need to offer realistic combustion rate responses to different engine characteristics such as port and fuel injector geometry. The current approach relies on a combination of empirical, phenomenological and experienced derived tools with poor accuracy outside the range of experimental data used to validate the tool chain, therefore making the exploration of unconventional solutions challenging. An alternative method that is less data and user experience dependent, is therefore needed to enable radical improvements in performance to be delivered without compromising the time to market. In this work, a pragmatic engine development process using a combination of a 0-D combustion Stochastic Reactor Model (SRM) provided by LOGESoft and non-combusting ?cold' CFD is used. The SRM captures the combustion chemistry in a computationally-efficient manner but does not capture in isolation geometric variables such as port and piston geometry. These effects are obtained via a CFD analysis which provides various inputs to the SRM to characterize the in-cylinder flow. Changes in Turbulent Kinetic Energy (k) and its dissipation (ϵ) in response to load and start of injection (SOI) have been investigated using CFD to develop a physically-based map for turbulent mixing time (t). This map-based approach reduced the number of cold CFD runs required, making the hybrid 0-D/3-D fast enough to deliver useful data in the early combustion system development phase of a new engine development program. Results have shown that a single baseline cold CFD run was sufficient to obtain a good correlation for the engine Rate of Heat Release (RoHR) and the knock tendency at the explored conditions. Further, a comparison of two injectors characterized by different sprays patterns has shown that the developed correlation correctly predicts the RoHR for two different tumble levels at the operating condition explored.