• 

  Syntra: Bold. Minimal. Intelligent.

Inspiration

Learning complex topics can often feel overwhelming and disconnected. The inspiration behind Syntra was to build an AI-driven platform that not only visualizes knowledge as an interactive graph, but also tracks progress and recommends personalized next steps. It aims to guide learners through their journey with intelligence and clarity.

What it does

Syntra transforms fragmented knowledge into a dynamic knowledge graph where concepts auto-link based on AI analysis. It aggregates multimedia and scholarly content from Wikipedia, YouTube, and arXiv. Users can explore subtopics, track their mastery levels, and receive AI-powered suggestions to bridge gaps and advance their learning.

How we built it

We built Syntra using React and Vite for the frontend, Node.js and Express for the backend, and Python for AI services. The backend connects to external APIs for content retrieval (Youtube, Wiki, Arxiv), uses Gemini AI to generate subtopics and connections, and leverages Snowflake and Cortex AI (Mistral AI) for tracking and personalized recommendations.

Challenges we ran into

Handling asynchronous multi-source data fetches proved challenging, particularly with slower APIs like arXiv. We implemented separate loading strategies and robust error handling to ensure frontend responsiveness. API integration with Snowflake was also difficult because it was my first time and I kept messing up making the DB and getting the right keys. Ensuring that the UI turned out good was quite difficult and time consuming. Scalability and real-time analytics integration with Snowflake also required significant effort.

Accomplishments

Successfully integrated AI-driven auto-linking of concepts and subtopics. Created a responsive and accessible knowledge graph UI that dynamically updates. Enabled mastery tracking with seamless real-time updates reflected in node visualization. Implemented a scalable backend architecture supporting data analytics and recommendations.

What we learned

We learned the importance of decoupling slow external data loads to improve UX, the critical role of AI in enhancing semantic connections, and how to design interactive tools for learning and using AI to aid in that process. Leveraging data warehousing like Snowflake can transform user interactions into insightful learning analytics.

What's next for Syntra

Next, we plan to integrate additional APIs such as Khan Academy and news article sources to enrich content diversity. As the user base grows, Cortex AI will analyze broader student success patterns and translate those insights to benefit all users, making Syntra a truly adaptive learning platform.

Built With

• fastapi
• gemini
• node.js
• python
• react
• snowflake
• vite