Aurora

• 

  For You Page

• 

  Look Up Page

What it does: Aurora provides users with research topic recommendations through an interactive swipe-based interface. It uses Gemini AI to generate topics tailored to user preferences and enriches them with metadata from Semantic Scholar, ensuring users receive high-quality, relevant research materials. Swiping right saves topics for further exploration, while swiping left dismisses them. How we built it:     •    Frontend: Developed using SwiftUI for a smooth, gesture-based experience.     •    AI Model: Utilized Gemini API to generate topic recommendations.     •    Research Data: Integrated with Semantic Scholar and arXiv APIs to fetch real-world research metadata based on AI-suggested topics.     •    Recommendation System: Designed a queue-based mechanism for fetching and displaying research topics dynamically. Challenges we ran into:     •    Parsing JSON responses from AI-generated content required refining extraction techniques.     •    Ensuring smooth UI interactions, especially with swipe animations and transitions.     •    Optimizing API calls to minimize delays while fetching research metadata.     •    Maintaining a balance between AI-generated topics and real-world research relevance. Accomplishments that we're proud of:     •    Successfully merging AI-generated recommendations with real research data.     •    Creating a visually appealing, intuitive swipeable interface.     •    Implementing a smooth and responsive user experience.     •    Overcoming API challenges to fetch and display relevant research metadata effectively. What we learned:     •    The importance of structuring AI prompts for optimal JSON responses.     •    Techniques for handling asynchronous data fetching in SwiftUI.     •    Enhancing user engagement through interactive UI components. What's next for Aurora:     •    Implementing a more advanced recommendation algorithm based on user interactions.     •    Allowing users to save and organize recommended topics into collections.     •    Adding deeper insights, such as citation trends and author networks.     •    Expanding the search functionality to allow direct topic exploration beyond recommendations.     •    Optimizing API calls for even faster metadata retrieval.

Built With

• apis
• arxiv
• gemini
• swift