LaTA: A Drop-in, FERPA-Compliant Local-LLM Autograder for Upper-Division STEM Coursework

LaTA is an open-source, FERPA-compliant local-LLM autograder for upper-division STEM courses, designed to run on-premises to mitigate data risk. It uses a four-stage pipeline and a local chain-of-thought LLM to grade LaTeX-native student work against reference solutions using a YAML rubric.