A DATA ENGINEERING AND DATA SCIENCE APPROACH TO STRENGTHENING CLOUD SECURITY THROUGH ML-BASED MFA AND DYNAMIC CRYPTOGRAPHY

Abstract

Cloud security faces increasing challenges due to large-scale data flows, evolving cyber threats, and the limitations of static authentication and cryptographic mechanisms. This research proposes an integrated Data Engineering and Data Science framework that enhances cloud security through Machine Learning–based Multi-Factor Authentication (MFA) and adaptive cryptography. The system employs engineered data pipelines to efficiently collect, preprocess, and analyze user behavior patterns using anomaly detection and predictive analytics models. Machine Learning algorithms dynamically evaluate authentication signals—such as device fingerprints, geolocation, access timing, and behavioral biometrics—to generate risk-aware MFA triggers. Concurrently, adaptive cryptographic techniques adjust encryption strength in real time based on threat intelligence and contextual risk scores derived from Data Science models. Experimental results demonstrate improved resistance to credential-based attacks, reduced false authentication attempts, and optimized cryptographic overhead. The proposed framework illustrates the critical role of Data Engineering and Data Science in enabling intelligent, scalable, and resilient cloud security architectures.