Smart Lecture APP

• 

  This screenshot showcasing the dynamic whiteboard and AI professor in action

The idea for the Smart Lecture APP stemmed from the need for more engaging and interactive educational tools. Traditional online learning methods often lack the personal touch and flexibility that students need. I wanted to create a system where students can interact with an AI professor in real-time, ask questions, and see complex concepts visually explained on a dynamic whiteboard.

What I learned Throughout the development of this project, I learned about Gemini Live API, real-time voice processing, and multimodal AI. I also gained hands-on experience with Google Cloud services, particularly Cloud Run, which allowed me to deploy the app with ease and scale it efficiently.

How I built the project The project was built using Next.js for the frontend, which interacts with the Node.js backend to handle WebSocket connections. The AI layer is powered by Gemini Live API, which processes real-time audio and generates visual commands to render on the whiteboard. The entire system is hosted on Google Cloud Run, ensuring scalability.

Challenges faced One of the biggest challenges I faced was ensuring that the real-time audio and visual synchronization was smooth, with minimal latency. Additionally, implementing the voice interaction and handling interruptions in the lecture was a complex task that required precise state management between the frontend and backend.

Built With

• express.js
• javascript
• next.js
• node.js