-
-
Classify Logo
-
Page for Class
-
Student Page
-
Login
-
Teacher Page
Inspiration
In a world where innovation breaks news every other day, people tend to ignore the uneasy fact that in 2024, only 33% of high school seniors were prepared for college-level math; only 30% of 8th graders were deemed proficient in reading; and, in 2025, a record low of only 35% of Americans were satisfied with our public education. Unlike virtually any other field, technology has made no substantial impact on education. Existing tools and platforms simply create a paperless environment, without necessarily enhancing the quality of education, nor lessening the burden on educators' shoulders. As past students of this great yet underserved system, we are inspired to create a platform where every student gets the attention of a private tutor, and every educator receives the help of an assistant analyst.
What it does
Classify is a virtual classroom service designed to be focused on shining a light on a student's process of work, instead of their results. Traditional platforms offering virtual classroom services only provide two discrete points - the empty worksheet and the completed copy. We believe the value of education is in the in-between. Through LLMs and hand-writing recognition, we process and analyze student work down to the step, question, and assignment level. This information is gathered and aggregated into several levels of reports for educators to understand each student's personal strengths and weaknesses. The report is also given to our Tavus API, creating a personalized, virtual tutor which students could reach out to 24/7. The educator also have the option to allow students to access real-time feedback as they work on assignments, creating a learning environment which fosters conceptual understanding rather than formulaic memorization.
How We Built It
Our team split into pairs of two, focusing separately on the frontend and backend. For the frontend, we used React and MathJax, optimizing the Classify experience for web. For the backend, we used a Python Flask server to handle requests and Postgress as our database.
Challenges We Ran Into
We ran into some bumps while working on the core necessities of our app: converting written work to text. We had never had to convert images to text in such a way that was so time sensitive, and we had to try many different solutions.
Accomplishments That We're Proud Of
We are proud of being able to provide powerful feedback to the educator, empowering them to understand more about each student with less. We are proud of creating an infinite working canvas, which allows students to work without fear of running out of space. We are proud of incorporating virtual tutors who actually understand the needs of each student based on real data. Aside from Classify's core functionalities, we are also proud of incorporating quality-of-life features such as the ability to convert a PDF directly into Classify questions, giving educators a seamless transition onto our system.
What We Learned
Throughout the process of developing Classify, we learned the importance of keeping your goal in mind throughout the development process. After every feature implemented, we asked ourselves, "how could this improve learning outcomes for students or improve the teaching process for teachers?" This allowed us to create a fully developed application, ready for scale, that best meets the needs of the core stakeholders in Edtech.
What's next for Classify
With our fully developed database, authentication, and fleshed out features, Classify is fully ready for launch. Our internal trials proved it to be a capable tool for K-12 and some university content. Our plan for Classify is to grow it within the public classrooms we once grew in and to empower the educators who once empowered us.
Log in or sign up for Devpost to join the conversation.