Summary: Assigning work to human resources is an important and common task for companies with high labour intensities. It is particularly complex if the human resources have multiple skills and if their efficiencies change due to learning and depreciation of knowledge depending on the tasks they are assigned to. In this context, up to now, only little work has considered strategic targets for company skill levels. Hence, we present an optimisation model to address the problem of assigning project work to multi-skilled internal and external human resources while considering learning, depreciation of knowledge and company skill level targets. This constrained nonlinear continuous model is solved using COIN-OR’s Ipopt. We assess the computation time and the impacts of problem parameters such as the shape of the learning curve, company skill level targets, depreciation of knowledge and cost parameters. Amongst other findings, our results show that faster learning of human resources leads to more specialised individuals while at the same time the whole company’s knowledge is broadened.