The growing popularity of multimodal resources in technology-mediated learning and teaching practices has brought to the fore the issue of learners' competencies in interpreting, employing and interacting with various semiotic resources, of which language is just one (Kress, van Leeuwen, Multimodal discourse: the modes and media of contemporary communication. Arnold, London, 2001). This is best captured by the concept of multimodal literacy, understood as participants' awareness of the affordances of available modes to enable 'transformative engagement' (Bezemer, Kress, Multimodality, learning and communication: a social semiotic frame. Routledge, London, 2016) with a wide range of meaning making resources/sign making systems. In this chapter we explore the meaning-making processes of learners participating in virtual exchange (VE) and using the available multimodal resources for the presentation of self. The core constructs that provide the background to our study come from the field of social semiotics. In line with Kress (Multimodality. A social semiotic approach to contemporary Communication. Routledge, London, 2010), we conceptualise online language learners and teachers who display semiotic awareness as autonomous sign makers representing their ideas through multiple modalities. As we argue, learners who are aware of semiotic resources of different modes and are capable of engaging with them in an informed, transformative way, exercise enhanced levels of autonomy. Throughout the chapter we demonstrate that a task-based approach to multimodal literacy development in VE setting can foster learners' autonomy in how they engage with available semiotic resources. [For the complete volume, "Language Education in Digital Spaces: Perspectives on Autonomy and Interaction. Educational Linguistics. Volume 52," see ED660763.]