UT Austin
Austin, TX
I had the opportunity to work as both a TA as well as a research assistant. My research focused on the intersection between NLP and software engineering.
As a TA for ECE 422C: Software Engineering II, I held office hours and led weekly discussion sections. I also helped grade exams and worked closely with rest of the instruction team to make sure the course ran smoothly.
I have been a research assistant in Dr. Milos Gligoric's lab for three semesters, where I worked on test completion using deep learning. Our paper on learning code semantics for test completion was accepted to ICSE'23.
Challenge
Writing tests is often seen as a time-consuming yet necessary facet of software engineering. In recent years, there has been a lot of attention on code completion, but not much on test completion using machine learning. We used code semantic data as input for the model and used test execution for rereanking generated tests.
My role
I assisted with a variety of tasks, from data processing to fine-tuning transformer models. I also helped write and edit the paper we submitted.
Technologies
- Python
- Pytorch
- TypeScript