Aceprep

Past paper content is spread across PDFs and sites. Students need feedback that matches official marking expectations.

Without structured metadata, tutoring cannot reliably tie explanations back to marks and examiner commentary.

We built a pipeline to collect and normalize papers, indexing by question structure, marks, commentary, and marking schemes.

A Q&A layer retrieves relevant segments and uses LLMs to answer or validate student work, grounded in schemes.

Students interact with a structured exam archive rather than unstructured PDFs.

Answers reflect examiner language and scheme logic.