ABSTRACT
In this chapter we provide fundamentals around building automated scoring engines for production (versus research) use. The focus is not solely on architectural design but also on key considerations influencing that design because good architecture arises from identifying stakeholder concerns and realistic use cases and successful systems built on that architecture require strong software development processes. Specifically, we first outline common stakeholder contexts and use cases specific to automated scoring implementations. Next, we describe software considerations that impact the design of automated scoring software and provide recommendations regarding architecture in key areas. Finally, we close the chapter with recommendations regarding team roles and structure.
