Over the past few years, learning object (LO) repositories have become valuable educational resources in a variety of instructional settings. However, there is a lack of detailed documentation about the actual process to be adopted under different conditions to create such a repository. The commonly established process adopts a synchronous model in which there is continuous face-to-face communication between the various members of the team -the subject matter experts, the instructional designers, the code developers and the reviewers. The synchronous process has resulted in the production of good quality LOs but in restricted domains and in small numbers. This model is not suitable to scale the LO production process up along numbers and across multiple domains without compromising on quality. In this paper we propose an asynchronous model for rapid, large-scale development of LOs in multiple science and engineering domains at the tertiary level of education. We document the problems encountered in adopting a synchronous model of LO production, present a comparative analysis of different asynchronous models, and detail out the methodology for the successful asynchronous model that we adopted.