AI for Threat Detection in Project Development

Abstract

This thesis presents the design and implementation of CloudNova, an AI-powered threat detection system developed to identify misconfigurations and potential security risks in cloud environments. The project integrates rule-based detection modules with a machine learning model to evaluate configuration files, logs, and cloud activity data. Using Python and the Flask framework, CloudNova analyzes uploaded inputs, predicts a security risk score through a trained Decision Tree Classifier, and generates structured reports in JSON and PDF formats. The system also includes an educational AI assistant that explains security concepts and a “CloudMind” generator that creates secure YAML configurations for services similar to AWS. Testing demonstrated that CloudNova effectively detects common misconfigurations and produces accurate risk evaluations in simulated environments. The results confirm that combining automation with artificial intelligence can enhance DevSecOps workflows, improve configuration security, and assist developers in preventing threats early in the project lifecycle. Future work includes integrating live cloud scanning, GPT-based assistance, and auto-remediation features.