The debate on digital labour platforms (DLPs) posits that algorithms engineer pervasive organizational control (Griesbach et al., 2019), while increasingly acknowledging that workers are capable of evading and resisting such control by manipulating algorithmic decisions (Kellogg et al., 2020). This second strand of research has significantly contributed to overcoming the deterministic assumptions that characterized early debates on the role of algorithms in social processes. However, by framing the interaction between algorithms and workers solely through the dialectic of control and resistance, it tends to overlook the variety of such interactions and their consequences at the individual and socio-organizational level.

Drawing on the concept of imbrication (Leonardi, 2013), this article seeks to move beyond this dichotomous framing by illuminating the complex and nuanced dynamics that characterize everyday interactions between workers and algorithms within food-delivery platforms, and by highlighting the heterogeneous outcomes these interactions produce.

Based on an ethnographic fieldwork conducted in Milan in 2020 - during which the author worked as a food-delivery courier for seven months - this study sheds light on the internal differentiation within the food-delivery workforce and analyzes how such differentiation is reproduced in practice, particularly through workers' varying capacities to interact with algorithmic technologies.

First, it compares two food-delivery platforms - Glovo and Deliveroo - to uncover how couriers are differently configured within their respective digital architectures, and how these configurations shape distinct forms of agency.

Second, it examines how two groups of workers, endowed with unequal cultural and socioeconomic resources, engage with and appropriate algorithmic systems in divergent ways.

The analysis reveals that pre-existing social stratification among workers is reproduced through processes of imbrication to platform, leading to the emergence of distinct working styles - namely, reactive and strategic - which, in turn, produce varying financial outcomes and different work subjectivities. For example, while reactive workers tend to accept any delivery assigned to them without question, strategic workers have developed a range of tactics aimed at manipulating order allocation in their favor or, more broadly, at selectively choosing the jobs they undertake. Although these strategic practices may appear to diverge from algorithmic decision-making, however, we conclude that they should not primarily be understood as acts of resistance to organizational control. Rather, they function as self-optimization strategies that are, to some extent, anticipated and accommodated by the food-delivery platforms themselves (Cini, 2025).