Automated Student Feedback

Hope you're well!
We’re a non-profit working with over 8,000 students across the UK, and we’re building a system to automate the marking and feedback process for student submissions on Google Classroom.

The system will need to:
- Download student work as PDFs from Google Classroom
- Feed each submission into an AI model to generate feedback and a grade
- Return that feedback to students
- Flag harmful content for manual review
- Highlight the top 10 submissions each week for staff review

We’re currently exploring two possible approaches:
- One using Google Classroom APIs and an online AI model (e.g. ChatGPT via API)
- Another using browser automation (e.g. Python + Selenium) with a locally hosted AI model

We’re open to suggestions on the best way forward and would love to hear your thoughts on what’s most technically viable and efficient. This is part of a wider effort to improve education and safeguarding — and while we’re a non-profit, we do have a solid budget set aside for this work.

If this sounds like something you could support, please do get in touch. We'd love to hear how you'd approach it.

Thanks!
Kai.