The utilization of metal additive manufacturing processes is still limited due to significant process condition variations, which cause defects and quality variations in the printed parts. The inability to predict and control variations in melting process conditions leads to qualification and certification challenges, hence hindering the wide adoption of these processes in industry. Despite vast investments in research and development of new control techniques in recent years, there remains no effective standard solution for fine-tuning process parameters to reduce undesirable variations in melting process conditions. The overall goal is to finely adjust process input parameters such as scan or tool moving speed and laser power along the scan/tool path in order to keep melting conditions in a desirable range. As one way, these input process parameters adjustments can be based on predictions from physics-model-based numerical simulations. However, for processes such as laser powder bed fusion of metals, currently there are no feasible solutions for performing high-resolution physics-based numerical simulations on a whole build scale. But recently, data-driven approaches to predict process conditions variations along the scan/tool path have been developed and validated in real experiments. These approaches can be used to significantly accelerate time-consuming physics-based numerical simulations to a point where it is feasible to predict process conditions variations along the scan/tool path for a whole build. The problem is that there is no single existing data-driven method that would be able to take into account all common sources of process conditions variations, i.e., the geometrical features of the observed part, the scan pattern, and the surrounding parts on the build plate. The combination of recently developed data-driven methods, which together enable prediction of process condition variations that source from all common sources of variations and which can be used to accelerate physics-based numerical simulations, will be introduced.