Tools
Tools
Kandadi, Akhila (2025) Scalable Threat Assessment Machine Learning Framework for CI/CD DevSecOps Pipelines. Masters thesis, Dublin, National College of Ireland.
Preview |
PDF (Master of Science)
Download (1MB) | Preview |
Preview |
PDF (Configuration Manual)
Download (790kB) | Preview |
Abstract
The current CI/CD pipelines using machine learning (ML) models to predict, detect, and respond to security risks are still, however, vulnerable to smart cybersecurity attacks that challenge the traditional tools used for security. These approaches to policing security systems create a hurdle to the deployment of code since they face the problem of overhead on developers caused by frequent false positive incidences. The security function is an essential part of the product in the DevSecOps culture. Moreover, enabling teams to be fully automated with the rapid growth of technology is a challenge. The hybrid transformer-Graph Neural Network (tGNN) ensemble proposed in this research is a groundbreaking model, focusing on the combination of the temporal pattern recognition offered by transformer networks with the structural relationship analysis of GNN that allows the detection of attacks in real-time. It is trained using the CICIDS-2017 dataset with class imbalance addressed through focal loss with log-dampened class weights and stratified sampling that preserves temporal coherence essential for sequential deep learning models. The hybrid tGNN ensemble achieves 99.98% accuracy with a 0.0005% false positive rate. The framework bridges the critical integration gap between advanced ML capabilities and practical DevSecOps requirements.
| Item Type: | Thesis (Masters) |
|---|---|
| Supervisors: | Name Email Gupta, Shaguna UNSPECIFIED |
| Subjects: | T Technology > T Technology (General) > Information Technology > Cloud computing Q Science > Q Science (General) > Self-organizing systems. Conscious automata > Machine learning |
| Divisions: | School of Computing > Master of Science in Cloud Computing |
| Depositing User: | Ciara O'Brien |
| Date Deposited: | 31 Aug 2026 15:14 |
| Last Modified: | 31 Aug 2026 15:14 |
| URI: | https://norma.ncirl.ie/id/eprint/9710 |
Actions (login required)
 |
View Item |