Abstract—Privacy-preserving analytics is essential in healthcare and financial domains where sensitive data must comply with regulations such as the General Data Protection Regulation (GDPR), the Health Insurance Portability and Accountability Act (HIPAA), and the Payment Card Industry Data Security Standard (PCI-DSS). Although Federated Learning (FL) enables collaborative model training without sharing raw data, existing frameworks often lack compliance-aware orchestration, efficient encrypted query execution, domain flexibility, and scalability under Homomorphic Encryption (HE). This paper presents FedPrivEngine, a federated and distributed data engineering framework designed for regulation-aware analytics across institutional silos. The framework integrates Homomorphic Encryption (HE), optional Differential Privacy (DP), Spark-based distributed orchestration, and a compliance-aware query planner to enforce policy constraints during federated execution. Two domain-specific models are implemented: HealthPrivNet, a Gated Recurrent Unit (GRU)-based model evaluated on the Open Safely healthcare dataset, and FinanceRiskNet, a Multi-Layer Perceptron (MLP)-based model assessed on the FICO HELOC credit risk dataset. Experimental results show that FedPrivEngine achieves high classification accuracy (93.8% in healthcare and 91.2% in finance), ensures zero policy violations, and supports encrypted query execution with success rates above 98%, while maintaining sublinear scalability and moderate system overhead. These results demonstrate the feasibility of secure, accurate, and regulation-aware federated analytics in policy-sensitive environments.
 
Keywords—Federated Learning (FL), privacy-preserving analytics, Homomorphic Encryption (HE), compliance-aware orchestration, healthcare and finance


Copyright © 2026 by the authors. This is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited (CC BY 4.0).