Application of Large Language Models in Software Development: Review of the Current State and Development Perspectives

Abstract

Large Language Models (LLMs) are rapidly transforming software engineering by automating tasks such as code generation, test creation, debugging, and documentation. Their integration across the software development lifecycle, from requirements analysis to deployment, introduces both unprecedented opportunities and serious challenges. This paper offers a comprehensive review of LLM applications in software engineering, contextualized within the frameworks of LLMOps and DevOps. It examines how specialized models and autonomous agents are reshaping workflows and proposes a framework for assessing organizational readiness for LLM adoption. In addition to technical capabilities, we discuss limitations such as hallucinations, context constraints, and the lack of explainability, as well as broader concerns around licensing, security, and developer deskilling. Drawing from recent research and tooling ecosystems, we synthesize the emerging best practices for safely deploying LLMs in production environments. This article aims to equip practitioners and researchers with a structured understanding of the current state and future direction of LLM-assisted software development.