This talk examines upstream assumptions about how to operationalize race in attempts to identify and address racial discrimination in sociotechnical systems. Recent work in algorithmic fairness has highlighted the challenge of defining racial categories for the purposes of anti-discrimination. Drawing on an analysis of papers published at the ACM Conference on Fairness, Accountability, and Transparency (FAccT), I describe how race is conceptualized and formalized in algorithmic fairness frameworks. I also discuss researcher reasoning behind different different operationationalizations of race and values associated with these choices, before reflecting on the institutional influences that shape racial classification practices. While categories used in algorithmic fairness work often echo legal frameworks, our analysis demonstrates that values from academic computer science play an equally important role in the construction of racial categories. Although the construction of racial categories is a value-laden process with significant social and political consequences for the project of algorithmic fairness, I argue that the widespread lack of justification around the operationalization of race reflects institutional norms that allow these political decisions to remain obscured within the backstage of knowledge production. Finally, I discuss wider implications about the role of assumptions in technical implementations of policy.