Volume 18, Number 2
From Pipelines to Platforms: The Impact of Platform-Centric Data Architecture on Enterprise AI in Regulated Industries
Authors
Mohammed Arbaaz Shareef, Anblicks, USA
Abstract
The article discusses the transition from pipeline-oriented data engineering to platform-centric data management as enterprises scale AI operations under regulatory constraints. The practical motivation stems from the rising complexity of heterogeneous sources, the growing demand for traceability, and operational friction caused by isolated ETL pipelines that fragment quality controls and accountability. Scientific novelty lies in aligning platform concepts (lakehouse unification, data products, mesh/fabric logics, and governed feature stores) with governance requirements for auditability and trustworthy AI lifecycle oversight. The article aims to analytically substantiate how platform-centric data design affects delivery speed, reuse economics, and compliance posture compared to pipeline-first operating models. The study relies on a comparative synthesis of recent peer-reviewed literature on open data platforms, data products, data fabric paradigms, and feature store architectures, complemented by governance research on data quality management and responsible AI governance. The conclusion formulates structured implications for regulated enterprises designing AI-ready data foundations.
Keywords
platform-centric data architecture, data platform, lakehouse, data products, data mesh