GradeMy.ai

Inspiration

As a college student, I often used ChatGPT to estimate my grades by simulating my professors’ grading styles. This helped me gauge my performance and identify areas for improvement. GradeMy.ai was born out of this idea, aiming to provide a reliable, personalized grading experience for students looking to optimize their assignments before submission.

What it does

GradeMy.ai allows students to upload their assignments and receive a simulated grade based on their professor’s grading preferences, course details, rigor level, specific assignment prompt, and rubrics (if available). Users can input additional notes, such as particular grading quirks (e.g., a preference for in-text citations or specific formatting styles). The app then evaluates the work, providing an estimated grade and actionable feedback to help improve the assignment.

How we built it

We built GradeMy.ai using a blend of AI-driven natural language processing and machine learning. The backend leverages a fine-tuned language model to interpret various grading styles and simulate feedback based on the given rubric and preferences. The front end was designed for ease of use, enabling students to input detailed information about their professor’s style and quickly receive feedback.

Challenges we ran into

One of the biggest challenges was ensuring the AI accurately replicated diverse grading preferences, from strict adherence to citation styles to subjective evaluation criteria like the “quote-sandwich method.” Creating a flexible, user-friendly interface for inputting these unique preferences and effectively training the model to interpret them posed significant hurdles.

Accomplishments that we're proud of

We're proud of developing an intuitive tool that students can rely on for meaningful, personalized feedback. Creating an adaptable grading simulator that handles the nuances of different grading styles and levels of rigor is a significant achievement. Our model’s accuracy in predicting grades and providing targeted feedback based on user input has been incredibly rewarding to see in action.

What we learned

Throughout development, we learned the importance of balancing simplicity with customization. Students benefit most from a tool that feels intuitive yet allows for detailed input when necessary. We also deepened our understanding of machine learning models' adaptability, as we continually improved the model's responsiveness to unique grading quirks.

What's next for GradeMy.ai

Moving forward, we aim to expand GradeMy.ai’s features, incorporating additional customization options and refining the grading accuracy across various academic disciplines. We plan to introduce a “revision history” feature, allowing users to track improvements over time and gather insights into their academic growth.

Built With

• javascript
• jsx
• jsx-**frontend-frameworks:**-react.js
• mongodb
• next.js
• openai-api
• react.js