Case Study
CampusNova
CampusNova turns static degree audits into clear, actionable advising workflows for students and advisors.
Founder & AI Engineer
2026
Problem
Degree audits are static PDFs disconnected from advising workflows, causing wasted meeting time and preventable graduation risk.
Solution
I built an end-to-end platform that parses DARS audits, structures requirement data, and generates personalized advising recommendations using a two-model AI pipeline.
Impact
Active development with a pending pilot in UTK's College of Emerging and Collaborative Studies; the core upload-to-personalized-advising loop runs reliably.
Problem
Degree audits hold the truth about graduation status, but they are static PDFs. Advisors waste meeting time reconstructing student standing, and students miss requirements too late.
Build
- PDF upload parses the DARS audit into completed, in-progress, and missing requirements.
- Requirement categories classify dynamically across major core, VolCore, capstone, and electives.
- Semester planner maps remaining courses across future terms.
- Recommendation layer suggests 3-5 course options based on remaining requirements and student goals.
- Advisor dashboard reviews submitted plans and full advising context.
Technical Signal
- Parsed unstructured DARS PDFs into normalized requirement state with PyMuPDF and Gemini.
- Built a two-model recommendation pipeline: Gemini grounds course-catalog matches, Claude writes the advisor-facing response.
- Added course cart, plan submission, advisor review, Supabase auth, and persistent student data.
- Designed the system around trust: prerequisite checks, in-progress exclusions, and reasoning surfaced with each recommendation.
Current State
Active development with a pending UTK CECS advising pilot. Core loop works: audit upload, structured requirements, personalized advising, and advisor review. Next: model evals, LLM-as-judge verification, and Canvas context.
Why It Matters
Trust is the product. One wrong course recommendation breaks the workflow. CampusNova surfaces why each suggestion satisfies a requirement, fits the student's goal, and belongs in the plan.