Co-authored during my undergraduate studies at Manipal Institute of Technology in 2023.
This paper proposes an architectural model for simplifying IoT system setup through a common management layer over sensor discovery, device provisioning, and cloud integration. It describes how these capabilities can work together to reduce device-specific configuration and support heterogeneous sensors.
A reference implementation using Raspberry Pi, CoAP, MQTT, and Azure IoT Hub demonstrates the approach, providing a foundation that developers can adapt when building IoT management frameworks.
Siamese Network Based Anomaly Detection Framework for Robust Federated Learning
Co-authored during my undergraduate studies at Manipal Institute of Technology and published in 2023.
This paper proposes a framework for protecting shared learning from malicious participants by separating update validation from model aggregation. It introduces a learned similarity-based approach to assessing client updates without inspecting their underlying training data.
A Siamese-network implementation demonstrates the approach on FEMNIST-derived models, providing a foundation for investigating computationally lightweight defenses against model poisoning in federated learning.
Full federated deployment remained future work.
Have a hard systems problem?
I am interested in the constraints, not just the technology.