Engineering intracellular metabolism has a long history, dated back to the late 1950s. The field has enjoyed significant attention in recent years, particularly because of the energy and environmental concerns. Many success stories have highlighted the feasibility of engineering intracellular metabolism to achieve a designed function. In almost all cases, the bottleneck of the problem resides on the biological side. With some exceptions, mathematical modeling has rarely played a predictive role in this development, mainly because of the lack of appropriate model structure and useful parameters. This talk will discuss a few examples of metabolic engineering and examine the role of mathematical modeling in these developments. In particular, we will present an en ensemble modeling technique that aims to narrow the gap been experiment and model prediction.