Research Software Engineer
AIEA Research Lab, UC Santa Cruz
Python · Tree-based Analysis · UCI Engine APIs · LLM APIs · RAG · Symbolic Logic
Overview
Software engineer and undergraduate researcher at the AIEA Research Lab, building production autograding infrastructure for graduate- and undergraduate-level machine learning courses, and conducting research on improving LLM accuracy and interpretability.
Project: XAI Autograder
- Built production autograding infrastructure serving students across multiple machine learning CS courses
- Achieved 95% student preference rate for the new guidance system in a pilot study versus the previous autograding system
- Developed a tree-based code analysis system in Python to evaluate correctness, behavior, and logic of student implementations in ML chess assignments
- Integrated external Universal Chess Interface (UCI) engine APIs to benchmark student eval and minimax implementations for accuracy and performance
- Adapted autograder for undergraduate coursework, accelerating assignment iteration, deployment, and testing
- Followed Agile/Scrum practices with weekly sprints, code reviews, and standups