Through colonialism, external forces can alter and shift social structures and practices. It causes trans-generational, often normalized, invisible, and profound marginalization of the collective identities of local and indigenous populations. Decolonization is the resisting and undoing of colonial impacts. It's the process of reforming a society's social, cultural, economic, and political structure to reflect more of the local and indigenous values than that of the colonial rulers. While sociotechnical systems can support the identity work of marginalized communities, social computing researchers have also discussed how these systems can cause harassment, exclusion, and other kinds of harm, impose values, and exhibit colonial impulses. While previous works in HCI and social computing have focused on several marginalized communities, there is a dearth of literature on the identity works of colonially marginalized communities using ICT platforms. My research contributes to understanding how colonialism marginalized people in the Global South across various dimensions of identity (e.g., race, gender, sexuality, religion, caste, nationality), how sociotechnical systems reinstate colonial structures and values, and how computing platforms both support and impede colonially marginalized communities' identity expression and performances. My dissertation examines Bengali ethnolinguistic communities' identity work, expression, and performances in ICT spaces, who have been marginalized by foreign powers for ages and are underrepresented in computing. Building on decolonial and postcolonial perspectives with a historicist sensibility, in this dissertation, my mixed-method empirical studies on various platforms highlight users' agency, the role of content moderation, algorithms, datasets, online communities, and their policies seek to understand how the previously colonized Bengali people decolonize their identities through these platforms and how the embedded coloniality of different components of these systems impede their identity work and expression. In the first two studies using trace ethnography, I sought to understand how people use the predominantly text-based Q&A site Bengali Quora to collaboratively reclaim narrative agency as a form of identity decolonization and how the governance approach of biased human moderators, collective surveillance, and algorithmic coloniality impedes that process. I built on these studies in terms of multimodality, ways of communication, and moderation approaches. Through semi-structured interviews with content creators on YouTube, where the interaction among users, contrary to the text-based collaborative website Quora, is multimodal and content creator-centric. This study explores YouTubers' motivation and strategies for making videos toward decolonial discourse and how they navigate different challenges in their work. For the final study, I explored the feasibility of automated content moderation by examining one of its common components, like NLP (e.g., sentiment analysis) tools, algorithms, and datasets. In doing so, my work altogether informs the broader social computing literature on identity, content moderation, fairness and bias, social justice, and ICT for development. [The dissertation citations contained here are published with the permission of ProQuest LLC. Further reproduction is prohibited without permission. Copies of dissertations may be obtained by Telephone (800) 1-800-521-0600. Web page: http://www.proquest.com.bibliotheek.ehb.be/en-US/products/dissertations/individuals.shtml.]