About AthenaPrep

Inspirations

All of us have navigated the complex admissions process for colleges. One of us, an immigrant whose parents knew nothing about the process, was especially perplexed while trying to apply senior year while balancing challenging coursework. Fortunately, he was able to afford a college counselor that could guide him through the process and alleviate his anxieties. Now, he gives back to his community by helping those younger than him through the arduous admissions process. However, he finds that he receives far more requests than he has the time to accommodate. Furthermore, many of the people seeking help with college admissions don’t come from a socioeconomic background that can financially support such a service which often runs in the thousands, if not tens of thousands, of dollars. Thus, he believed he could make an impact in his community by creating an application that guides those through the process of college counseling at a fraction of the price.

AthenaPrep is a low-cost, effective counseling service powered by AI. Currently, it will help high school students reflect on their personal experiences and plan an effective common application essay. Furthermore, it will allow them to receive granular, personalized feedback on the essay drafts they create. In essence, they will be guided through one of the most difficult parts of the application. In the future, we hope to expand this to cater to the various other parts of the application as well. By leveraging technology for social good, we hope to bridge gaps in education equity, making the college application journey less overwhelming and more accessible for all.

What it does

AthenaPrep has two functions, Orientation: Uses an LLM to create a back and fourth conversation with a student, that creates thought provoking questions making connections with details provided, to narrow down a compelling essay topic. After 12 questions, the AI creates more final questions to provide closure, and then provide a detailed closing statement stating good approaches for a solid essay the student could consider. Essay Editor: Provides rich and detailed feedback to a student provided essay, focusing on general broad topics like cohenrency, and qualities that would be appealing to an admissions officer. Feedback is shown in the style of google docs comments, with text boxes shown beside the essay, and relevant text being highlighted.

How We Built It

AthenaPrep was built using a combination of technologies in both the backend and frontend, that work seamlessly together.

Tech Stack:

  • Frontend: React, TailwindCSS, Vite
  • Backend: OpenAI, Flask, Python, Gunicorn
  • AI Integration: We used a large language model (LLM) to power the essay feedback feature, specifically ChatGPT 4o.

Our goal was to create a platform that was intuitive for students to use. On the frontend, we focused on building a clean and modern interface with TailwindCSS for responsive design, while React and Vite ensured fast load times and a dynamic user experience. The backend was designed mostly to handle the student responses, and do the API calls and processing that the project needed.

Challenges We Faced

Building AthenaPrep was not without its challenges.

  • AI Integration: One of the main challenges was integrating the large language model into the platform in a way that was seamless, but also had effective responses.

  • Frontend: With most of the team using vite and tailwind for the first time, there were hurdles we constantly ran into while learning and adapting to the tech.

  • Balancing Feedback: One of the largest challenges was properly tuning the prompts of the LLMs to generate exactly what we were hoping for. This required practically endless iterating and trial and error to create prompts that both worked properly, and had responses that felt actually effective and useful for the user.

Accomplishments We’re Proud Of

We are incredibly proud of the current state of our product, especially considering the short time frame that we implemented it. However, the truly impressive aspect is the effectiveness that was immediately seen in the models once they were tuned properly.

What We Learned

I feel the most valuable thing we learned was a completely new approach to designing an application, and deploying it. From Vite, React, Tailwind, Node.js, Flask, and many more frameworks/libraries that we integrated, we found an efficient approach to create something that both is practical, and pretty.

In the backend, we learned alot about using OpenAI's API, and ensuring the prompts formatted in the right way to be forwarded properly to the frontend.

In the front end, we learned a myriad of things we can use when designing future websites, to create something flashy yet minimalistic.

Learning how to plan together, using concepts like MVP and other strategies, we also learned how to coordinate in a way that allows everyone to be working at their best at any moment. This also included having great practice pulling and handling file conflicts with git. Being able to define a vision, and work towards it once we were on the same page taught us the value of quick thinking, and communication.

What's Next?

We’re incredibly proud of how far AthenaPrep has come, but we know this is just the beginning. Moving forward, we plan to continue refining the platform, by further optimizing and training the AI's feedback, finding ways to reduce operational costs, and adding new features to expand the scope of AthenaPrep.

College prep is an extremely broad category, and while AthenaPrep does work to counteract the effects of wealth-inequality on the admissions process, there is countless other features which could improve our impact as we strive to make the access to higher education equal for all.

Conclusion

At AthenaPrep, we’ve set out to build more than just a tool—we’re building a movement toward fairness and accessibility in education. By leveraging the power of AI, we’re helping students take control of their futures and navigate the college admissions process with confidence.

Built With

Share this project:

Updates