HYBRID DEEP LEARNING MODELS FOR SECURITY DEFECT DETECTION IN ENTERPRISE SOFTWARE SYSTEMS. ijiecit 2025, 1 (2), 15-21.