Inspiration

We're drowning in information, but starving for understanding. Every day, we ask AI thousands of questions and get instant answers from ChatGPT, Google, and other tools. But here's the problem: we forget almost everything we read. Furthermore, most people can't explain ideas clearly: Education students freeze during teaching practice Office workers stumble through presentations Everyone struggles to share ideas simply

Research shows that when you teach something, you remember 90% of it. When you just read it, you remember only 10%. This isn't new - it's called the Feynman Technique and it's backed by decades of learning science (Edgar Dale's Cone of Experience).

The famous physicist Richard Feynman said: "If you can't explain something simply, you don't really understand it."

That's exactly what we built: an app where you learn by teaching an AI student.

Instead of asking AI for answers, you explain concepts to AI. And the AI asks you questions back - just like a real student would. This forces your brain to truly understand, not just memorize.

What it does

User upload documents so that AI can be specific on the questions Based on the documents, the system will use AI to generate different kind of topics For each topic, user can fire a chat session to give explanations about that topic. Meanwhile, an AI model will constantly hear the user’s explanation and ask follow-up questions After the teaching-session finishes, our AI model will also give feedbacks: stating the strength and weaknesses ⇒ suggest ways to improve on the explanation User can view the teaching-session history and choose to re-explain a topic. Each attempt will have its own version stored in the DB

How we built it

  • We used Next.js for frontend because of it's standalone power when integrating with real-time AI models.
  • For backend, we used Nest.js & MongoDB for security and scale.
  • For real-time speech-to-speech sessions, we used the gpt-realtime-1.5 model
  • For evaluations and other stuff, we used GPT 4o mini model, with cheap price and high intelligence

Challenges we ran into

This was our first time participating in a large-scale hackathon, and honestly, we felt quite overwhelmed at the beginning. Seeing so many talented hackers and realizing the scale of the event made us both nervous and excited at the same time. However, as the event went on, we gradually adapted to the environment and even started to truly enjoy the experience.

It was also our first time working together under such tight time pressure. Because of that, miscommunication was inevitable. After splitting tasks among team members, we sometimes faced difficulties when merging our work, especially when conflicts occurred between branches. In addition, certain tasks depended on others being completed first, which occasionally slowed down our progress and made time management more challenging.

On top of that, the intense pressure took a toll on one of our team members, who experienced moments of mental exhaustion during the process. Despite these challenges, we supported each other, stayed determined, and ultimately managed to complete our project. Going through these difficulties together not only strengthened our teamwork but also made the accomplishment feel even more meaningful.

Accomplishments that we're proud of

Talking about our accomplishments, we have so much we’d love to share. First of all, of course, it’s our beloved website. We spent countless hours building it, pouring our effort, focus, and passion into every detail. Whether we receive an award or not, this project is still something we are incredibly proud of because it represents our hard work and dedication. We still remember how, on the first day, we only slept for about two hours just to keep developing and improving the website. It was exhausting, but also unforgettable. Needless to say, we truly love what we have created together.

What we learned

By joining this event, we gained an incredible amount of valuable experience from workshops hosted by OpenRouter, Agora, TinyFish, AWS, and many others. We were deeply inspired by the speakers, not only by their knowledge but also by their passion and genuine desire to create meaningful impact for the community. Their energy and dedication motivated us greatly. We truly appreciate the opportunity to learn from such inspiring individuals, and we hope that one day we will have the chance to sit in the audience again, listening, learning, and continuing to grow.

What's next for ReverseLearningAI

  • Release Alpha version to the Vietnam's market.
  • Implement more advanced features, like deep research on a keyword to find many different kind of topic to practice.

Built With

Share this project:

Updates