In job markets, there is evidence that particular pairs of firms frequently see workers changing jobs between them, while at the same time other firm pairs see little or no worker changes. This observation has led to the introduction of empirically determined Labor-Flow Networks, firm-to-firm networks that act as the substrate for job mobility. In this talk, I present the first model for the dynamics of the workforce in the labor flow network, and adjust it for two countries (Finland and Mexico) where large scale microdata is available for individuals and their employers. I find that the model correctly predicts the steady state of the job market, provides a new way to predict firm-size distributions based on microscopic mechanisms, and inspires a new notion of firm-specific unemployment, a concept that is likely to prove very helpful in understanding the detailed dynamics of the workforce.