Evaluating Student Worksheets at Scale with ADEOS Intelligence

A tutor conducts 1:1 online classes for students across the globe. At the end of every chapter, students are given a worksheet to assess their understanding. When the tutor evaluates each worksheet, it takes 30 to 40 minutes, hindering the tutor's time taken for planning and preparing for class and feedback sessions.

• No standard evaluation criteria; it varies by board, grade, subject, and chapters for students with respective worksheet structure.
• Handwritten worksheets submitted at low scan quality, so the tutor spends extra time reviewing or asking the student to reshare the page before correction
• Evaluation time scales linearly with student to worksheet volume, takes the tutor time to evaluate instead of interacting and understanding the students well.

The platform uses ADEOS to extract handwritten and printed student responses from scanned multi-page worksheets, allowing the grading engine to evaluate each answer against board-specific rubric.

• Enhances scanned pages to accurately read and recognize student responses across both OCR and vision-model pipelines.
• Analyzes and adapts to each board's dynamic worksheet structure by grade and subject.
• Evaluates MCQ, long-form, diagram-supported, and multi-step answers using a library of approved board–grade–subject rubrics, providing step-level partial marks to ensure consistent grading across the cohort.
• Drafts the full evaluation for the tutor to review, adjust, or reset. The tutor gives final approval before results reach the student.
• Delivers the evaluated worksheet to students and overall grades through LMS integration