GrAIdscope is a revolutionary AI-powered grading platform designed to transform the way educators assess student work.
In today’s fast-paced educational environment:
- Teachers are overwhelmed by grading large volumes of exams.
- Students often wait weeks for meaningful feedback.
GrAIdscope automates free-response exam evaluation with speed and precision:
- Teachers upload their answer key PDF.
- Students submit completed exams.
- AI compares submissions against the key and produces comprehensive grades and detailed, actionable feedback—all within minutes.
💡 Benefits:
- Consistency: Eliminates human error and subjective bias.
- Efficiency: Reduces grading time from weeks to minutes.
- Learning: Students receive immediate, transparent feedback.
For investors, GrAIdscope represents a cutting-edge opportunity at the intersection of AI, EdTech, and workflow automation.
We identified major issues with OSU’s current grading platform, Gradescope:
- ⏳ Slow: Manual grading in big classes takes 2+ weeks.
⚠️ Unreliable: Reports of lost student tests, missing grades, and software glitches.
We researched other solutions, compared their features, and designed a system that combines AI efficiency with reliability and accessibility.
| Feature | Description |
|---|---|
| Teacher Upload | Teachers upload the exam answer key PDF. |
| Student Submission | Students upload completed exam PDFs. |
| AI Grading | Compares student answers with teacher key and produces: • Detailed scoring breakdowns (Completeness, Correctness, Simplification/Presentation) • Overall exam score • Feedback comments per question |
| Secure & Fast | Uses API keys and backend processing to handle multiple submissions efficiently. |
| User-Friendly Frontend | Intuitive interface for uploading files and viewing results. |
Click to expand the design process
- Existing graders fail at handling handwritten text.
- Many AI grading tools have a steep learning curve for teachers.
- Tested multiple OCR engines, AI scoring logic, and dataset preprocessing.
- Brainstormed intuitive UI layout for students and teachers.
- Integrated OpenAI GPT-5 API for semantic grading.
- Used Amazon S3 for scalable file storage.
- React + Vite frontend for responsiveness.
- Google OAuth for authentication.
- Flask + PostgreSQL backend with NGINX deployment.
Outcome: Improved usability, data flow, OCR precision, and AI grading accuracy.
- ⚡ Grading Speed: Reduced midterm grading from 2 weeks to minutes.
- 🎯 Accuracy: AI scores closely match human graders with detailed explanations.
- 📈 Scalability: Handles hundreds of submissions concurrently.
- Web Server: NGINX
- Cloud Storage: Amazon S3
- Web Framework: Flask
- Database: PostgreSQL
- AI/ML Integration: OpenAI GPT-5 API
- Frameworks/Libraries: React, Vite
- Languages: HTML, CSS, JavaScript
git clone https://github.com/25kgozon/HackOHIO25.git
cd HackOHIO25