IoT-FedMalDetect

Federated Learning based Malware Detection for IoT Edge Devices.

This project introduces a novel semi-supervised Federated Learning framework for detecting malware on IoT devices without requiring labeled data at client nodes. The approach achieved a 37% reduction in training time and 20% lower memory usage compared to standard models. A sandboxed IoT testbed with simulated botnet attacks was used to validate the framework’s effectiveness and real-world resilience.

iGuard

Efficient Isolation Forest Design for Malicious Traffic Detection in Programmable Switches.

This project introduces iGuard, a novel unsupervised malicious traffic detection system optimized for deployment in programmable switch. By fusing the lightweight structure of Isolation Forests with knowledge distilled from autoencoders, iGuard achieves significantly improved detection accuracy—up to 48.3% higher—while maintaining a comparable or lower memory footprint. Implemented in P4 and evaluated on Intel Tofino switches, iGuard demonstrates real-time performance and practical deployability at line rate.

Recognition for Teaching Assistant

Honored by the Department of Computer Science & Engineering, IIT Hyderabad in recognition of my dedication and invaluable contributions as a Teaching Assistant in the Network Wireless Security Bootcamp and Faculty Updation Program 2025 (ISEA Phase 3).