An AI instructor-rating system that analyzes quantitative and qualitative data — course satisfaction, completion, content quality, learning outcomes — to grade instructors. The data surfaces what a single star rating hid.
If a 4.8-star average is the end of the conversation, that's not evaluation — it's under-evaluation.
Ratings skew toward satisfied users and compress all the quality of teaching into one average. Without separating course satisfaction, completion, content quality and learning outcomes, the operator's feedback to instructors stays at 'keep your ratings up.'
And instructors who don't see 'where to improve' don't change much from one course to the next.
We analyzed quantitative and qualitative signals across multiple axes, then returned not just a grade but 'where to strengthen' to each instructor.
Course satisfaction, completion, content quality and learning outcomes each have their own model. They aggregate into an S/A/B grade — not an absolute score, but a position relative to similar-domain instructors.
Instructors receive their grade alongside customized improvement insights, so the next course knows where to start, on data.
Separate models for quantitative and qualitative signals.
Data ingest, processing and grade computation on Spring Boot.
Operator admin and instructor-report UI on React.
Integrated to the operator's existing MySQL + AWS environment.