The rapid evolution of cyber threats has necessitated the adoption of advanced machine learning (ML) techniques for real-time threat detection and response. However, conventional ML-based cybersecurity systems rely heavily on centralized data collection, which raises significant privacy concerns, including data breaches, unauthorized access, and non-compliance with data protection regulations such as GDPR and the Digital Personal Data Protection (DPDP) Act. To address these challenges, this paper proposes a privacy-aware federated machine learning model designed for modern cybersecurity applications. The proposed framework leverages federated learning to enable decentralized model training across distributed client nodes, ensuring that sensitive data remains local and is never directly shared. To further strengthen privacy guarantees, differential privacy mechanisms are incorporated during local model updates, preventing inference attacks and model inversion risks. Secure aggregation techniques are employed to combine client updates into a global model while preserving confidentiality. The model is evaluated on benchmark intrusion detection datasets, including NSL-KDD and CICIDS 2017, using performance metrics such as accuracy, precision, recall, F1-score, privacy overhead, and communication cost. Experimental results demonstrate that the proposed approach achieves competitive detection performance while significantly enhancing data privacy and regulatory compliance.