<p>The rise of data-centric artificial intelligence (AI) has exposed a persistent misalignment between modern data-intensive practices and the structure of computing education, where databases, machine learning (ML), and scalable data systems are commonly taught as independent components. Although prior surveys address AI literacy, data science education, or Big Data infrastructures separately, they do not explain how competencies, pedagogy, assessment, and technological infrastructure must operate together within end-to-end data pipelines. This survey aims to address this gap. Based on a structured analysis of peer-reviewed literature published since 2020, this study synthesizes how contemporary curricula can be aligned with the engineering requirements of data-centric AI systems. The analysis adopts an engineering-oriented perspective that treats robustness, reproducibility, scalability, and system integration as foundational properties that shape educational design. Core data-centric competencies are identified and explicitly linked to learning objectives, instructional strategies, assessment models, and supporting technological systems in order to enable coherent pipeline-level reasoning. The survey further analyzes representative institutional deployments from academic and industrial contexts to illustrate how these dimensions are instantiated in practice and to delineate their current limitations. By separating conceptual, pedagogical, and infrastructural concerns while preserving their functional alignment, this study provides a modular basis for curriculum design in data-centric computing education. It clarifies what students are expected to learn, how instruction is structured, which systems support it, and how learning outcomes are evaluated within integrated data-centric curricula.</p>