Biography

is an expert in Civil Engineering with over 17 years of experience. He holds a Ph.D. in Structural Engineering. He has published numerous research papers in reputed journals. He has guided several research scholars. He has worked on infrastructure development projects. His research interests include structural analysis and sustainable construction. He has received awards for academic excellence. He contributes to curriculum development. He collaborates with industries on engineering projects. He is a member of professional bodies. He actively mentors students and researchers. He is dedicated to sustainable engineering practices.ficity. In a recent paper, the presented Smart trust, a hybrid deep learning framework for real-time cloud security threat detection that blends CNNs[13], Transformer models, and Long Short-Term Memory (LSTM) networks with Reinforcement Learning (RL). This framework is based on Zero-Trust Architecture (ZTA), a contemporary security paradigm that enforces least privilege norms by continuously assessing context and assuming no trust by default. In order to guarantee the accuracy and transparency[14] of security events and automated incident response procedures, Smart trust additionally integrates blockchain-based logging. When compared to traditional and hybrid baselines, Smart Trust considerably reduced false positive rates while achieving excellent detection accuracy (≈98–99%) across several threat types on benchmark datasets like CIC-IoT 2023 and UNSW-NB15[15]. This hybrid integration shows how deep learning can significantly enhance threat detection in complicated cloud environments, especially those supporting smart city services, when combined with adaptive decision systems (RL) and immutable logging (blockchain). Federated learning and other distributed learning paradigms have also been investigated for security in decentralized IoT and smart installations. These systems frequently combine local and global learning[16] to address privacy and scalability issues, even though they are not necessarily strictly hybrid. By cooperatively developing a shared intrusion detection model without centralized data pooling, federated techniques using ensemble knowledge distillation, for instance, have demonstrated efficacy in diverse IoT environments, enhancing privacy and lowering communication overhead. Blockchain-federated learning frameworks, which combine decentralized learning with consensus techniques to protect privacy while guaranteeing cooperative model upgrades across dispersed nodes, have also been proposed for adaptive threat protection[17]. A forward-thinking method for real-time cloud security threat detection in smart cities is represented by hybrid learning models that include several machine and deep learning approaches, frequently with adaptive elements like reinforcement learning. These models overcome several drawbacks of conventional IDS and static by utilizing the complementing strengths of various learning algorithms and contextual mechanisms like blockchain logging and Zero-Trust frameworks[18]. But striking a balance between interpretability, scalability, and accuracy is still a research issue. Securing the upcoming generation of smart city infrastructures will require ongoing research on federated, distributed, and explainable hybrid architectures.