Profile

Rupesh Raj Karn is a PostDoc Researcher in Electrical and Computer Engineering at the Khalifa University of Science and Technology, Abu Dhabi, UAE. He pursued his PhD at Khalifa University under supervision of Prof. Ibrahim Elfadel. His PhD thesis title was "Machine Learning Methods for the Automated Management of Cloud Computing Workloads". During his PhD, he worked as an intern for three months and research scholar for six months at IBM TJ Watson Research center, New York USA under the supervision of Dr. Prabhakar Kudva. During his tenure at IBM, he worked on the application of AI technologies to dynamic cloud environments and produced some important results, some of which have been published and others are in review. Specifically, he worked on building predictive autoscaling model to trace the cloud resource usage and scale them based on workload’s performance. He has created a DevOps architecture for distributed learning with dynamic model selection to handle the dynamic cloud monitoring signals. He built a framework for congestion detection in deep neural networks for multi-task progressive learning with application to cloud network analytics. Such a multi-task learning model was grown and pruned as needed to develop a compact neural network. To connect cloud and artificial intelligence, he built a setup using a well-known neuro-evolution algorithm and microservice evolutionary techniques and tested on cloud signals to perform analytics such as microservice signatures regarding OS, applications, behavior, and intrusion.