Inspiration

Working as an academic therapist for the last two years, Geetanjali noticed gaps between therapy and at-home learning. She observed:

1) Parents pay $120+ per hour for therapy it's expensive, and long-term most can’t afford it.

2) Writing observation notes and debriefing parents so the child can continue building skills after sessions is tedious; we need one place that bridges to at-home learning.

3) Many kids with academic disorders like dyscalculia have co-occurring neurodevelopmental disorders, making therapy hard to follow in time-bound sessions they need something engaging.

Each child is different, so strategies must be adapted to each child’s learning curve.

Hence, a therapist’s observations are gold combined with AI, a comic-based storytelling approach, and drills disguised as games, ALL in one app.

What it does

Our application uses guidance from experts in each type of disorder ( such as dyslexia (reading), dysgraphia (writing), and dyscalculia (math) ) and combines this guidance with exemplary literature to create our learning exercises for the child. This is complimented with our suggestions for parents to notice behavior patterns for the child to enable them to learn beyond just our platform. This holds the potential to give the child and the parent a happier life with lesser issues as they deal with the learning disorder together.

Our adaptive learning system guides parents through taking inputs on the child's behaviors and learns from the child's performance in our in app exercises. We use Lang Chain and RAG to adapt our exercises and responses by learning from the parents' feedback. Our app understands the child’s journey with more nuance and creates a track to help them learn faster resulting in a more normal lifestyle.

Our application collates all the information and creates a dashboard for the caregiver to visualize their growth which is critical because this is only validated by a therapist and costs a fortune.

Challenges we ran into

The AI model was too big to upload to github. So, we had to zip the source code with AI model.

Built With

Share this project:

Updates