Interactive Physics Learning Platform
Simulare is an interactive physics learning platform that combines textbook-based information retrieval with visual simulations and voice interaction. It addresses a common student problem: physics concepts like the Coriolis force are difficult to visualize through textbooks alone. Simulare bridges this gap by creating a multisensory educational experience grounded in verified academic sources, using a RAG pipeline to prevent AI hallucinations.
View ProjectSimulare is an interactive physics learning platform that combines textbook-based information retrieval with visual simulations and voice interaction. It addresses a common student problem: physics concepts like the Coriolis force are difficult to visualize through textbooks alone. Simulare bridges this gap by creating a multisensory educational experience grounded in verified academic sources, using a RAG pipeline to prevent AI hallucinations.
Development Time
Team Members
Awards Won
Platforms Supported
Textbook-grounded physics explanations retrieved from undergraduate materials using a Snowflake-backed RAG pipeline, ensuring accuracy and preventing AI hallucinations.
AI-generated physics simulations powered by Google Gemini that users can customize and regenerate to explore concepts like the Coriolis force visually.
Integrated ElevenLabs voice interface for interactive discussions, guided explanations, and quizzes to reinforce learning through audio.
Built with React Native and Expo for seamless deployment across iOS, Android, macOS, and Windows with responsive design and cross-device sync.
Adapting Snowflake's API beyond standard Streamlit implementations for a custom RAG pipeline with physics textbook data.
Managing ElevenLabs dependency conflicts with React Native UI components, causing build and runtime errors.
Ensuring consistent performance and UI behavior across iOS, Android, macOS, and Windows simultaneously.
Built a custom integration layer to query the Snowflake corpus directly from the Express.js backend, enabling flexible retrieval-augmented generation.
Isolated the voice agent into a dedicated module with careful dependency management to prevent conflicts with the main UI layer.
Leveraged Expo Go and NativeWind CSS for platform-agnostic styling and thorough testing on each target platform.