Inspiration

Learning new skills often feels overwhelming-too many resources, no clear order, and no sense of progress. We wanted to make learning structured, visual, and optimised. In our own experience, we love learning new skills but sometimes even tools such as ChatGPT fail to create a coherient and precise plan to each our goals (it has resources that sometimes don't work, vague ideas and concepts displayed and too many resources linked at once). NeuraLearn aims to combat this.

What it does

NeuraLearn generates AI-powered knowledge graphs for any topic and computes the most efficient personalized learning path, helping users always know what to learn next.

How we built it

We built a full-stack app using Python and Flask for the backend, React and TypeScript for the frontend, and graph algorithms like Ant Colony Optimisation and spectral analysis for pathfinding.

Challenges we ran into

Designing accurate topic dependencies, tuning optimisation algorithms, and balancing performance with visualisation complexity were major technical challenges.

Accomplishments that we're proud of

We successfully combined advanced graph algorithms with a polished UI to create an end-to-end intelligent learning system within hackathon time constraints.

What we learned

We gained experience applying theoretical algorithms to real problems, integrating AI with full-stack systems, and designing for user motivation and clarity.

What's next for Neuralearn

We plan to add adaptive learning based on user performance, better source validation, collaboration features, and deployment at scale. Other things we plan to improve include the overall UI elements and optimisation of the overall site and algorithms (we care currently using a gnn we made using torch and though its effective at creating a difficulty metre, we can to make the site more faster as its taking around 5 seconds to build a graph as well as the gnn indications and metrics).

Built With

Share this project:

Updates