RocketGrades

Inspiration

The inspiration behind this project stems from our members own frustrations we have experienced as students waiting for assignment grades to be posted. RocketGrades solves this issue by providing a source of nearly instant feedback for both professors and students.

What it does

RocketGrades is a Web Application that serves as a Learning Management System (LMS) similar to platforms such as Canvas and Blackboard. RocketGrades provides an easy-to-use interface in which professors and students can view materials regarding coursework, classes, grades, and more. Both user types are presented with a respective dashboard that neatly organizes the core functionalities necessary for their roles. RocketGrades innovates upon virtual learning environments such as Canvas in that it enables professors to automatically obtain feedback on submitted assignments through the use of the Mistral API, thus solving the issue of prolonged uncertainty in regards to academic standing.

How we built it

Frontend

Angular - We used Angular for its component-based structure which allowed us to quickly build an interface that was dynamic and responsive. Its modular architecture allowed us to maintain an organized codebase throughout the entire development process.

TailwindCSS - TailwindCSS provided us a highly-customizable and rapid prototyping capabilities for user interface designing. The use of daisyUI, a tailwind plugin, further contributed to a clean interface designed for high usability without sacrificing development and project performance.

Auth0 - Auth0 was implemented in RocketGrades for user sign-up and authentication, providing teachers and students a secure application experience.

Backend

Node.js - Node.js served as the backbone of our server-side logic, allowing us to handle asynchronous operations and API integrations efficiently.

Mistral AI -

La Plateforme AI by Mistral was implemented in our Web App as a chat bot that communicates with the class and student details database to give relevant feedback based on assignment submissions using LangChain query routing.
MistralAI with it's structured outputs feature that supported JSON on LangChain was used to grade the assignment submissions.

MongoDB Atlas- We implemented MongoDB due to its use of a NoSQL database which gave us flexibility due to its unstructured data formatting for quick formatting.

AWS S3 - Amazon Web Services S3 was used as storage for our files which supported our ai features. We also attempted to implement a cloud backend through the use of elastic beanstalk, while the node backend and it's routes run well, the chat requests seem to break it.

Challenges we ran into

AWS Services Integration - Angular struggled to interact and deploy global variables to the browser which the browser required for establishing a connection using AWS.

Auth0 Setup - It was difficult to setup in Angular due to the requirement of assigning metadata to either a teacher or a student role. It was also implemented after a non-authenticated login was already created for the application.
better secret management Secrets could've been managed between different features of the app. there is still an aws secret in the git repo (we deactivated it before we made the repo public)

Accomplishments that we're proud of

Real Functionality - Our team is proud of being able to develop and implement real functionality and infrastructure for RocketGrades. It is worth noting that we could have easily simulated functionality through the use of mock data but our sights were set on a application with non-simulated functionality.

Diverse Technology Integration - We are also proud of the wide variety of services, tools, and technologies we utilized in the development of our web application. By using an extensive tech-stack, we took on a more complex and difficult challenge that was more rewarding to make progress on.