Developed a model for the Databricks Assistant LLM to classify the relevance of input context data, improving response accuracy and reducing input tokens by as much as 22%.

Working on compilers & systems research with Prof. Michael Franz at the UC Irvine Secure Systems & Software Lab.

Worked on designing a recommendation system to use our NLP and WatsonX AI models to intelligently suggest fixes for server faults detected with Instana, our application performance monitoring product, with the goal of eventually using AI to fix some types of errors automatically without human intervention.

Invited as one of a small group of students to visit Jane Street in NYC for the opportunity to learn about quantitative finance and how Jane Street applies CS & Math during a program for select first and second-year undergraduates.

- Led creation of CPI's popular competitive programming classes program, which has since benefited 1,000+ students and netted substantial donations for the Competitive Programming Initiative.- Designed Usaco Guide's Groups feature, a LMS that we use in our classes to help aggregate all information for students in one place and allow students to learn more effectively, with integrated automatic practice problem execution & grading.