ApplicationOpen applicationAIEducationOCR

Lingkee AI · Exercise and Marking Assistant

Help teachers prepare differentiated exercises within a preset Hong Kong Primary 5 Chinese unit, then review answer images, scores and feedback in a separate marking workflow.

Audience
Teachers & students

Screenshots & demos

Scroll to explore · Click to enlarge

Project overview

For Hong Kong Primary 5 Chinese teaching, this workspace brings exercise preparation and answer-image marking together. Exercise generation currently uses a preset lesson unit. Teachers provide a reference question within that scope or choose one from the available question bank, then select a question type and difficulty. Questions, reference answers and marking guidance are presented together so teachers can judge their fit with the lesson objectives.

Preparation: from reference question to exercise preview

  1. Find and adjust a reference: Search the question bank by type and keyword, preview the question, options, supporting material and available answer information, then bring it into the editor for adjustment.
  2. Choose question type and difficulty: Generation supports fill-in-the-blank, true-or-false, multiple-choice and longer questions. Standard and lower-difficulty modes provide starting points for different practice needs. The lower-difficulty direction favours shorter sentences, clearer guidance and fewer distractions; teachers still judge the language and curriculum suitability.
  3. Review the exercise preview: The exercise preview shows each question, its marks, options and answer or marking guidance, together with the question count and total marks. Teachers can review the learning focus, wording and scoring expectations rather than receiving a list of questions without supporting material.

Generated exercises are preparation aids. Exercise generation and marking remain separate workflows: a generated exercise does not automatically become the marking reference on the other page.

Marking: check the content before scoring

  • Answer image and recognised content: Marking starts by uploading or dropping in an image of a student's answers. The original image and recognized content provide side-by-side references, including question text, responses, blank answers and items needing confirmation.
  • Additional marking context: Teachers can repeat recognition or add lesson background, question requirements and marking notes before starting scoring separately.
  • Question-by-question feedback: The report presents a total score, individual marks, corresponding feedback and an overall comment to support discussion of particular misunderstandings and subsequent practice.

Teacher review and material requirements

The staged process gives teachers an opportunity to compare the image, recognized text and scores before providing feedback. PDFs need to be converted to images first. Sample questions, answers and scores demonstrate the workflow and are not results from the current material. During actual use, teachers should check that the content matches the student's original work and confirm the final questions and scores.

Ownership & delivery

I independently owned the teaching workflow, product interactions, AI orchestration, frontend, backend and deployment. Built distinct question-generation and marking flows with question-bank search, exercise previews, answer recognition and item-level feedback, keeping teacher review and grading context central to the experience.

Explore more work with similar capabilities or uses.

ApplicationOpen application

AI Mock Interview

Prepare questions from a target role and personal experience, then practise with spoken questions, typed or dictated answers, follow-ups and detailed feedback on what to work on next.

AIEducationBusiness Tool
View case study