CET Student Affairs Funding

We will build an AI-assisted autograding agent for UVU’s core CS courses (CS 1410/2420). The agent runs student programs in a secure sandbox, checks them against instructor tests (unit, hidden stress, mutation), and turns failures into concise, human-readable hints—nudging, not spoiling. Canvas integration returns grades and feedback within minute,s so students iterate while concepts are fresh. The immediate beneficiaries are 120–200 students per semester, including first-generation learners who often avoid office hours yet respond to timely, low-stakes feedback. Faculty and TAs gain time back from repetitive grading to focus on code reviews and mentoring. We will openly share containers, test templates, and evaluation rubrics so other UVU courses—and peer institutions—can replicate the workflow. Success looks like faster time-to-first feedback (<15 minutes for most submissions), more submissions per assignment, higher median correctness, and a documented, reusable playbook for responsible AI in grading.