Public research funding programs typically make grants with minimal intervention by program staff, rather than using a hands-on approach to project management, which is more common in the private sector. In contrast, program staff at the US Department of Energy's Advanced Research Projects Agency Energy (ARPA-E) are given a set of real options with which to manage funded projects: abandon, contract or expand project budgets or timelines. Using internal data from ARPA-E, we show that active project management enables risk mitigation across a portfolio of research projects. We find that program staff modify projects frequently, especially project timelines, and these changes are more sensitive to poor performance than to strong performance. We also find that projects with a shortened timeline or reduced budget are less likely to generate short-term research outputs, compared to those of ultimately similar size. This evidence suggests that the practice of active project management, when combined with high upfront risk tolerance, can be used to enhance the productivity of mission-oriented public research funding.