■Finding Ace coders in GitHub: Designed a software recruitment tool to find candidates from 3TB, 36Million Users GitHub data set with specific language skills and location criterion.■ Orchestrated ETL pipeline using a spark cluster and Postgres SQL with Web front end. ■ Optimized PySpark SQL queries to derive and update metrics to process the data set in 3hrs mins. The response time on the front end is 300ms■ Tools : AWS EC2, S3, Spark SQL, Dash, Plotly

■ Designed Intelligent systems for edge analytics specifically executing real-time applications. ■ Lead a project to design ContextAiDe, a middle ware and API to execute real time Mobile Crowd Sensing. ContextAiDe helps to design real-time apps like disaster monitoring, real time event capture, finding missing child, tracking a perpetrator using context optimized recruitment and predictive analytics for various device context. In case of perpetrator tracking example, it results in savings in energy (37.8%), time (43%) and enhanced user experience by obtaining context relevant data.■ Team player, collaborator in design of an automated VM management scheme to optimize data center operations running collocated LTE base station applications. Developed an extension to OpenStack scheduler for elastic design resulting in resources predicted with an accuracy of 85%. ■ Lead the Intel project of medical control applications for real-time patient monitoring using infusion control to maintain a specific drug level in the patient’s body e.g. blood glucose. Enabled parallel processing to explore architecture parallelism to meet real time constraints of 10 s by performance evaluation and mapping for monitoring multiple patients. ■ Collaborator in Smart Grid project that enable a cloud-based Linear State Estimator and False Data Detection which obtains Phasor Measurement Unit (PMU) data from smart grid and process it in time constraints 23 ms for up to 1500 nodes.■ Lead predictive analytics to capture the different seasonality patterns of data center parameters such as electricity prices, solar and wind, workload. This predictions are used for optimization of various data center parameters to efficiently executed workload or for designing peak shaving using energy storage devices.■ Teaching Assistant to 500+ students for courses of Mobile Computing, Digital Design of Circuits, Principal of Programming Languages, Big data computing course.

■ Developed design of an automated VM management scheme to optimize data center operations running collocated LTE base station applications. This scheduler predicts the usage pattern of VMs to consolidate and migrate VMs, allowing scaling of the VMs with increase in usage. Developed Python extension module to OpenStack scheduler for elastic design predicting the resources with an accuracy of 85%.■ Lead the Intel project of medical control applications for real-time patient monitoring using infusion control to maintain a specific drug level in the patient’s body e.g. blood glucose. Enabled parallel processing to explore architecture parallelism of Intel Xeon Phi, GPU and i7 multi core devices to meet real time constraints of 10s by performance evaluation and mapping for monitoring multiple patients.

■ Design a Disaster Recovery Strategy and Continuity Planning System. Researched resiliency, fault-tolerant aspects for growing data center applications. Designed a decision tree-based framework for the Disaster Recovery Levels in an existing data center.

■ Lead a team of 13 teaching assistant to organize and functioning in the ELET labs in College of Technology. ■ Included teaching over 200 students in the semiconductors devices and circuit laboratory, introductions of labs, including students discussions, helping performing the lab assignments. ■ Coordinating the labs to progress in accordance with the theory lectures.■ Organization meeting, tasks and handling issues on everyday basis.