We discuss one aspect of the allocation of new scientists to teaching and research careers. In the past, this allocation problem was treated on the basis of a model with known constant parameters and as a classical open-loop optimal control problem with the allocation ratio as the sole control variable. The utility function in that treatment took both short term and long term goals into account. Here we allow for uncertainty in the possibly time-varying system parameters, and we account for the possibility of new scientists going into careers other than teaching and research. We treat the allocation problem not as an optimal control one but rather as one robust control, insensitive to uncertainties, in order to assure desired numbers of teachers and scientists within a computable horizon.