A CASE STUDY ON INTELLIGENT AUTOMATION FOR CI/CD OPTIMIZATION IN CLOUD-NATIVE ENVIRONMENTS

Abstract

In recent years, Continuous Integration/Deployment (CI/CD) has become an indispensable practice in software
engineering. These practices automate DevOps work, speed up software delivery and maintain quality by minimizing human error.
This paper explores the use of intelligent automation and machine learning to improve the performance, reliability, and
predictability of cloud-based CI/CD pipelines. Due to human-induced bottlenecks, long build times, and low visibility, traditional
pipelines are not suitable for supporting modern software systems that require fast, error-free releases. To overcome these issues,
the study incorporates AI-based methods, e.g., predictive analytics, anomaly recognition, and autopilot-based decision-making,
into CI/CD processes. The research framework proposed above uses data collection, data preprocessing, feature engineering, and
Support Vector Machine (SVM) modeling to forecast failures and maximize resource utilization. Case-based analysis shows that
with the introduction of ML, there are significant performance gains, including shorter build times, fewer deployment failures,
reduced resource use, and increased model accuracy. It has been experimentally validated that ML-enhanced CI/CD pipelines can
reduce build time by 33%, failure rate by 60%, and achieve significant improvements in precision, recall, and F1-score. This
publication demonstrates the use of AI-based DevOps to provide a multi-cloud experience characterized by intelligent, scalable,
and proactive software delivery.