PROJECT // 2024
AUTOMATED GRADING INFRASTRUCTURE
CMU CS Software Developer (Contract)
Discourse — Methodology
Architected the backend for a departmental assignment grading system serving 100+ students. Implemented Docker-based code execution sandboxes and integrated GPT-3.5 APIs for automated feedback generation on code quality.
Concl. — Scientific Impact
Reduced grading latency while maintaining feedback consistency via containerized execution and LLM-based analysis.
§