Authors: Michael Reed, Abigail Foster, Victoria Stewart, Chaitanya Srinivas, Rishi Kumar
Abstract: Platform engineering has emerged as a critical discipline for managing the growing com-plexity of modern cloud-native ecosystems, enabling organizations to standardize infra-structure, streamline software delivery, and improve operational efficiency. The rapid ad-vancement of Large Language Models (LLMs) has introduced new opportunities for trans-forming platform engineering through intelligent automation, natural language–driven inter-actions, automated decision support, and enhanced operational intelligence. This study presents an evidence mapping analysis of the application of Large Language Models in platform engineering, with the objective of identifying current research trends, implementa-tion approaches, technological capabilities, benefits, challenges, and future development directions. The study systematically examines existing literature across key domains includ-ing infrastructure automation, DevOps, Site Reliability Engineering (SRE), AIOps, cloud operations, incident management, observability, and software delivery pipelines. The find-ings indicate that LLM-driven solutions significantly enhance platform engineering by au-tomating routine operational tasks, accelerating troubleshooting processes, improving knowledge management, supporting intelligent workflow orchestration, and enabling con-text-aware operational decision-making. Furthermore, the integration of generative artificial intelligence with cloud-native platforms, infrastructure-as-code frameworks, and autono-mous operations is creating new possibilities for self-managing and adaptive digital envi-ronments. The evidence mapping also identifies important research challenges related to model reliability, explainability, security, governance, privacy, hallucination mitigation, and responsible AI adoption in enterprise operations. The study concludes that Large Language Models represent a transformative force in the evolution of platform engineering and are expected to play a central role in the development of intelligent, autonomous, and highly resilient operational ecosystems, providing significant opportunities for future research and innovation.
DOI: http://doi.org/10.5281/zenodo.20630453
