This research investigates how machine learning models can identify malicious activity within encrypted traffic flows without decrypting sensitive user data. The project emphasizes privacy-aware threat detection techniques suitable for enterprise and cloud-scale security monitoring.