AI-Driven Nutrition Assistant for Thalassemia Care

Inspiration

Thalassemia patients in India face several interconnected challenges, including chronic micronutrient deficiencies, a risk of iron overload, and emotional stress. Existing nutritional guidance is often not culturally relevant or linguistically accessible. The inspiration was to create a comprehensive tool that addresses these unique problems within the Indian context.

What it does

The AI-Driven Nutrition Assistant for Thalassemia Care is a mobile app that offers personalized and culturally intelligent support. It tracks diet and mood, provides food recommendations based on Indian cuisine and regional preferences, and includes an emotional support chat. The app also sends medication and chelation reminders, and it allows for monitoring by doctors and caregivers.

How we built it

The app was built using a specific technology stack. The frontend was developed with React Native, Tailwind or shadcn/ui, and Framer Motion. The backend uses Node.js or Python (FastAPI) with Supabase/PostgreSQL for user data and food logs. AI modules include GPT-4/BERT for emotional chat and mood detection, and a rule-based engine for iron warnings and meal suggestions. The app also integrates a database of Indian cuisine and uses n8n for multilingual support in Hindi, Tamil, and Bengali. Integrations include Google Fit, Fitbit, and Lab APIs.

Challenges we ran into

We faced several challenges during development. These included collecting data on regional meals and nutrient profiles at scale, and ensuring the accuracy of iron-related recommendations without clinical testing. Another challenge was balancing sensitive emotional interactions through AI and dealing with the complexity of multilingual NLP for mood and food terms. We also had to consider internet access gaps in rural areas and the need for an offline fallback.

Accomplishments that we're proud of

We are proud of our ability to create a solution that blends nutrition science, emotional intelligence, and the diversity of Indian cuisine. We successfully prototyped an interactive UI with features for food and mood logging and a basic medication reminder system during the hackathon.

What we learned

We learned that addressing chronic illnesses requires a comprehensive approach that considers not just the clinical aspects, but also behavioral, emotional, and cultural factors. We also learned the importance of creating a scalable solution for communities and rural outreach.

What's next for AI-Driven Nutrition Assistant for Thalassemia Care

Following the hackathon, our next steps include launching a nutritionist/doctor portal for remote monitoring and integrating lab reports for personalized dietary adjustments. We plan to expand the regional food database and train multilingual NLP models. We will also build a personalization engine based on user behavior and implement smart nudging AI using reinforcement learning. Finally, we will conduct clinical validation and pilot programs with NGOs.

Built With

• bert
• google
• gpt-4
• n8n
• node.js
• react-native
• supabase
• tailwind