HokieGPT
Inspiration
Our journey began with a vision to create a PDF-based AI assistant specifically designed for VT students. As we progressed, it became evident that each student has unique learning styles and preferences. This insight inspired us to expand our mission: to develop an AI that delivers personalized learning experiences, maximizing both motivation and efficiency by catering to individual needs.
What it does
HokieGPT analyzes the content of uploaded PDFs and intelligently responds to user questions. More than just an AI chatbot, it tailors its responses and resources based on each user's learning preferences, generating custom "DEI Markdown"—a feature that personalizes content to support diverse learning styles and promote inclusivity.
How we built it
Our development process involved careful planning and extensive research into cutting-edge technologies. We explored multiple approaches, collaborated on implementation strategies, and seamlessly integrated features to ensure the AI's effectiveness in delivering personalized learning experiences.
Tech Stack
- Next.js App Router: For the core framework, ensuring a smooth and efficient user experience.
- OpenAI: To power embeddings, enabling the AI to understand and process user queries.
- LangChain.js: For Retrieval-Augmented Generation (RAG) functionality, making the AI responses more context-aware.
- MongoDB Atlas: For storing and managing vectorized data, providing robust and scalable database support.
- Bytescale: For handling PDF storage efficiently and securely.
- Clerk: For user authentication, ensuring secure access control.
- Tailwind CSS: For styling, enabling rapid and responsive UI design.
Challenges we ran into
Integrating the frontend and backend proved to be our biggest hurdle. Ensuring smooth communication between the two was critical to creating a cohesive user experience, and this required both time and problem-solving.
Accomplishments that we're proud of
We’re proud to have successfully built a functional PDF-based chat AI that not only works but also exceeds our initial expectations. The DEI Markdown feature is particularly noteworthy, as it allows for truly personalized content creation that aligns with diverse learning needs.
What we learned
We gained invaluable lessons on the importance of efficient time management, especially when working under the pressure of a hackathon. Strategic planning and prioritization were key to our success.
What's next for HokieGPT
We have exciting features in the pipeline, including a Markdown display function, text-to-speech capability, interactive quizzes, and more. Our vision is to continually evolve HokieGPT into a robust, all-in-one learning assistant for every student.
Log in or sign up for Devpost to join the conversation.