CultureClass: Personalized Learning Paths

"GPT + Qloo = AI-curated learning plans from your personal culture."

Comment

Inspiration

I realized that people learn best when they’re emotionally invested, and what better way to spark interest than through someone’s favorite movie, music, food, or book? With GPT-4 and Qloo’s Taste AI, I saw an opportunity to bridge culture and education by generating highly personalized learning journeys. That’s how CultureClass was born.

What it does

CultureClass takes a user’s favorite: 🎬 Movie 🎧 Music artist 🍽️ Food 📚 Book 💡 Topic they want to learn ...and uses AI to generate a personalized step-by-step learning path, complete with actionable resources and recommendations (like Spotify links, restaurants, gear, or books).

How I built it

Qloo was super helpful for this. It cleaned up user input, added cultural context, and made GPT’s suggestions way more relevant. It basically acts like a taste-to-context engine, translating your personal preferences into something GPT can actually work with.

Frontend: React + TailwindCSS for a beautiful, responsive UI Backend: Node.js + Express API AI: OpenAI GPT-4 to generate personalized learning paths and classify the content Qloo API: Used to incorporate Taste AI to enhance recommendation context

Challenges I ran into

Building CultureClass solo in a short time was challenging. One of the biggest hurdles was learning how to properly integrate two powerful APIs Qloo for cultural data and GPT-4 for generating content, while keeping the backend clean and responsive. I also had to balance creativity with practicality, making sure the recommendations felt personalized but still worked technically. Debugging async data flow between the frontend and backend took a lot of trial and error. And since I was handling everything alone, from UI to API design to demo video, time management was a constant challenge. But in the end, it was worth it.

Accomplishments that I AM proud of

I'm proud that I built a fully functional end-to-end AI app from scratch, solo, within the hackathon deadline. Integrating Qloo’s Taste AI with GPT-4 to create truly personalized learning paths was a big win and seeing it work felt amazing.

What I learned

Building CultureClass gave me a deeper appreciation for how personalized experiences can dramatically boost learning engagement. I learned how to structure prompts for GPT-4 to generate clear, concise, and actionable learning steps based on a user's cultural preferences. Integrating the Qloo Taste AI API taught me the importance of refining user input, turning raw preferences into meaningful, AI-friendly context. I also strengthened my backend development skills by handling real-time API communication, formatting issues, and asynchronous data flow. As a solo developer, I had to balance feature scope with feasibility, which helped me focus on delivering a clean MVP that communicates the core value of blending culture with education. This project showed me how powerful AI can be when it meets thoughtful design.

What's next for CultureClass: Personalized Learning Paths

In the next phase, I aim to take CultureClass beyond generating a learning path by integrating real-world, actionable recommendations. Imagine finishing a step about coding and instantly getting links to the best beginner-friendly books, or learning about music theory and being directed to curated Spotify playlists. I also plan to connect APIs like Spotify, Google Books, Amazon, and even restaurant finders for food-related suggestions. The idea is to blur the line between learning and doing turning a static learning path into an interactive, resource-packed experience. The next version of CultureClass goes beyond static learning paths:

✅ Want to learn about coding? You’ll get links to beginner-friendly books and interactive tutorials ✅ Interested in music theory? Instantly access curated Spotify playlists ✅ Food lover exploring Japanese culture? Get local restaurant recommendations near

I didn’t implement these features in this version due to two main reasons: since I am doing this solo, I had time constraints and, along with that, budget limitations. Integrating multiple APIs requires not only proper backend handling but also paid tiers for services like Amazon product search or advanced Spotify integrations. As a recent graduate, I had to prioritize building a functional, end-to-end MVP that clearly demonstrates the core idea blending culture and learning through AI without overcomplicating the initial build. This hackathon version is focused on proving the concept, but the vision for the next iteration is much bigger and actionable.

Built With

Share this project:

Updates