Teaching Assistant for Dr. Yale Patt and Dr. Ramesh Yerraballi for the course EE 306: Introduction to Computing, and EE 319K: Introduction to Embedded Systems.I am responsible for holding recitation and review sessions for students, as well as office hours to assist them. I also prepare and grade exams, homework and study resources.

Created a real-time debugging tool for chip performance using Elasticsearch, Logstash and Kibana. The tool creates meaningful visualizations and detailed statistics about important events like exceptions, interrupts, etc. Wrote a pre-cursor ELK API wrapper.Tested and debugged AVX-512 instructions.Improved the Jenkins Bugreport Integration for auto-updates.Created a landing page in React for all frontend links on Traefik.Improved SystemVerilog code for register diff checking.

Developed a deep learning model in Keras to detect and eject polypropylene in fast-flowing industrial cotton. Used the pre-trained VGG16 model with ImageNet weights.Achieved a total accuracy of 96% with a 100% accuracy on true positives. Worked with very minimal training data and highly imbalanced classes (about a 1:50 positive to negative ratio).

Worked on iCount - a commercial product used to count seat occupancy in theaters across the country. The camera would take a picture of the audience in the theater, and the software would then detect which seats are empty and which are occupied using a machine learning model. I reworked the machine learning model in Tensorflow/Keras using the pre-trained state of the art Resnet50 model (with ImageNet weights). Achieved an accuracy of 87% with minimal training data.Through this project, I learnt deep learning and Keras over the course of 3 months.