In this paper, we propose a model for an energy broker who acts as a third party between the grid and its clients, through the maintenance of a two-sided portfolio of bilateral contracts. These contracts are positioned for roles involving a finite number of pointwise interventions within specified availability periods, and are well suited for clients such as distributed generators or ancillary services, on the grid side. The setting is otherwise quite general and may for instance involve bilateral contracts on wholesale markets. While bilateral contracting hedges the parties involved against the volatility of energy prices, the management of the broker portfolio raises a number of modelling and computational issues, which stem from the aggregation of disparate resources. To address these challenges, we devise an innovative algorithmic framework that involves robust optimization with respect to short term decisions, factoring in long term information obtained from a secondary model that embeds the full extent of all contracts' durations.