RedCell Guardian

Precision AI tools to aid diagnosis, care, and awareness

Inspiration

Thalassemia patients face lifelong challenges — from diagnosis delays to transfusion dependency and lack of awareness in rural communities. Inspired by this gap, we envisioned an AI-powered tool that can assist in early detection, improve patient education, and support ongoing care — especially for under-resourced areas.

What it does

RedCell Guardian is a modular AI tool that offers:

Thalassemia risk prediction using basic blood parameters like MCV, MCH, Hb, RBC count
Chatbot assistant that answers FAQs related to Thalassemia symptoms, treatment, and lifestyle
Localized awareness content in English and regional languages (text and voice)
Optional: Alerts for transfusion schedules or medication tracking
Optional: Donor–recipient match suggestions based on user profiles

How we built it

ML Models: Python with scikit-learn (decision trees, logistic regression)
Data: Synthetic samples + publicly available medical datasets
Frontend: Streamlit (for web), Jetpack Compose (if mobile)
Backend: Supabase / Firebase
Chatbot: Hugging Face or Rasa
Other Tools: Figma (UI mockups), GitHub (collaboration)

Challenges we ran into

Sourcing real and diverse datasets for Thalassemia risk modeling
Designing a model that gives interpretable results for non-medical users
Building an interface that is friendly for both caregivers and patients
Balancing the scope — avoiding feature overload within a short timeframe

Accomplishments that we're proud of

Ideating a real-world impact solution grounded in medical relevance
Planning a design that is inclusive, localized, and user-friendly
Establishing a clear roadmap for rapid prototyping and testing
Creating an early model prototype with rule-based prediction logic

What we learned

How to design AI systems that prioritize ethics, interpretability, and inclusivity
The importance of accessible UI/UX in health tech for diverse users
How to break down a complex healthcare challenge into modular tech components
That even small AI tools can have meaningful impact when thoughtfully built

What's next for RedCell Guardian

Collect medical feedback to refine model inputs and thresholds
Build a working MVP with prediction, chatbot, and awareness modules
Translate UI content into 2–3 Indian languages
Explore collaboration with NGOs or blood banks for pilot testing
Open-source the tool and invite contributions from the health-AI community