systems researcher and programmer, open source contributor and an academic writer.
Sep 2013 to Sep 2015
- Helped physically build the cluster from ground up.
- Assist with the operation and maintenance of the university’s 100 node HPC GPU cluster using SLURM.
- Identified and reduced performance overhead by 25% in the cluster caused by improper job scheduling.
- Implemented the initial prototype of running OpenStack on the HPC cluster.
Sep 2012 to May 2013
- Virtualized 20% of the physical servers to run on a private cloud, cutting 50% of the costs on power.
- Reduced the down time for several web/mail servers up to 50% that occurred after system update by writing scripts to automate database initialization and automated the procedure for Operating System patches and updates.
Jan 2011 to May 2012
- Algorithms and Data Structures (Graduate level)
- Network Security (Graduate level)
- Discrete Mathematics (Undergraduate level)
since Jan 2009
- Conducted extensive experiments on Cloud Platforms such as AWS, Azure and Google Cloud.
- Improved the performance of the Virtual Machine's backing up (checkpointing) mechanism by 70%.
- Implemented a standalone disk-storage simulator from scratch.
- Developed models to analyze terabytes of storage trace data in a sub-millisecond interval.
- Migrated batch jobs from Virtual Machines to Docker on AWS.
- Implemented modules that run on VM Manager to reduce performance bottlenecks.
- Published several projects on top-tier conferences and Journals.
- Implemented a novel security mechanism for Cloud Systems.
- Implemented a framework for online detection of attacks in Virtual Machines in AWS.
- Used memory introspection techniques to protect VMs from memory based attacks.
- Implemented a Hadoop Job scheduler for customized load balance.
- Modified the Linux CPU scheduler to avoid interference between VMs.
- Implemented a new priority level built into Xen Hypervisor Credit Scheduler.
- Implemented a tool to study the resource utilization of 192 production servers.
- Developed power-aware software systems that utilized 10% less power in the data center.
- Implemented a system that efficiently mined Enterprise search engine corpus.
- Assisted in developing privacy preserving location (Anonymous GPS) based systems.
- Was one among 12 students to receive the first ever NSF I-Corps award to carry the research at Stanford University.
- Implemented a new workload based performance attack of Cloud Systems.
- Conducted extensive experimentation of the attack on Amazon Web Services platform.
- Implemented several algorithms to increase the CPU and memory utilization of the Cloud.
- Proposed and tested models to run Hadoop jobs using data centers' spare resources efficiently.
- Developed power-aware software systems that utilized 10% less power compared to other similar systems.
- Implemented a system that efficiently mined Enterprise search engine corpus by issuing aggregate queries to its database
- Received NSF I-Corps award to carry the research at Stanford University.
- Developed mathematical models and algorithmic techniques for output compression, input pruning, and search space pruning.
- For search space pruning, I developed novel anti-monotonic properties to filter out candidate groups.
- Implemented statistically sound techniques to find skyline groups.
Ph.D. in Computer Science
The George Washington University, 2017
Certificate of Entrepreneurship
Stanford University, 2012
M.S. in Computer Security and Information Assurance
The George Washington University, 2011
M.S. in Computer Science
The George Washington University, 2010
B.E. in Computer Science
Anna University, 2004 - 2008