Metric Computation Infrastructure

Sketch Recognition Lab

• Integrating acceleration algorithms on VCK190 AI Core / FPGA using Vitis HLS, C++, and OpenCL which led to less power consumption, and faster execution time compared to CPU execution.

• Analyzing large experimental data sets using Python and Pandas.

• Led initiative among interns with Stanford ML Lab to collect Solar Panel data.

• Built data pipelining tools using Python and AWS that are now a part of NRELs open source libraries.

• Used AWS to pipeline data into large scale classification models.

• Technologies: AWS, AWS Systems Manager, Linux, Python, React.js

• Produced REST APIs and React.js dashboard to monitor performance of 1000+ client servers, significantly reducing downtime and streamlining fault detection, allowing developers to focus on core tasks instead of IT.

• Developed 30+ Java lambda functions for load balancing and monitoring EC2 instances.

• Utilized AWS and CloudWatch APIs to monitor and detect abnormal server activity, setting up custom metrics and alarms to identify and mitigate DOS attacks.