HaemoAI

HaemoAI: 60 seconds to save a life, AI for Thalassemia Free Tomorrow. Hope runs on Time, so should Blood, never let a patient die waiting for Blood.

Comment

Inspiration

India is prone to higher Thalassemia rates ; a condition that requires lifelong blood transfusions. Yet, families still wait hours, often desperately, just to find a compatible donor.

We were inspired by NGOs like Blood Warriors, who are doing incredible work on the ground with camps and fundraisers. But they are limited by manual systems ; WhatsApp groups, Excel sheets, no real coordination or intelligence to speed things up or automate .

We saw an opportunity to reimagine the process using AI, geolocation, and ethical data systems. Our idea: a platform that can match patients to donors in under 60 seconds and also get a good doctor consultation end to end .

What it does

HaemoAI is a concept for an AI powered, real time blood donation and Thalassemia care platform that puts speed, accuracy, and trust at its core.

Here’s how the platform would function in the real world:

  • A caregiver submits a blood request using a mobile app or chatbot
  • AI identifies the most likely eligible donors nearby, based on location and donation history
  • Donors receive notifications and can confirm availability instantly
  • The system updates the patient, hospital, and caregiver in real time
  • All donations are logged securely using blockchain to prevent misuse or duplicate entries
  • After donation, the patient receives follow up support and the donor gets impact feedback and recognition

The idea is to connect lives, not just match blood.

How we built the idea

We designed this solution by mapping real world problems faced by patients, donors, and NGOs, and linking them to scalable tech solutions.

We considered using:

  • Firebase and GeoFirestore for location-aware real-time donor matching
  • LSTM models to predict future donor availability
  • A multilingual chatbot (based on LLaMA 2) to support caregivers in multiple languages
  • Polygon blockchain for expiry logging and audit trails
  • Zero Knowledge Proofs and IPFS for secure health data sharing
  • A modular UI that can work as an app, web portal, or even SMS/voice bot for low connectivity areas

While we didn’t build a working prototype, we ensured that every layer of this system is both technically feasible and ethically sound.

Challenges we explored

As we built the idea, we ran into some deep design questions:

  • How do we make real time matching reliable in areas with low connectivity?
  • How do we preserve medical privacy while still sharing enough for donors to be verified?
  • What happens if a donor agrees but can’t show up? Who gets notified next?
  • How do we keep blockchain useful without slowing down emergency response?
  • How do we simplify the onboarding process for people unfamiliar with tech?

We created fallback plans, hybrid workflows, and human in the loop interventions to ensure reliability even without high tech infrastructure.

Accomplishments that we're proud of

  • We created an end to end idea that blends AI, healthcare, blockchain, and behavioral science
  • We made sure it’s inclusive, privacy first, and affordable for real deployment
  • We grounded our solution in real stories from the field, not just hypotheticals
  • We identified tech partners, APIs, cost models, and regulations that would support implementation
  • Most of all, we turned a complicated challenge into a simple, clear path to saving lives faster

What we learned

  • In blood donation, distance = delay = danger. A 5 km donor match can save a life
  • Tech alone isn’t enough, trust, community, and ethics matter more
  • AI can improve healthcare only when it's interpretable, inclusive, and aligned with policy
  • Gamification and personalized recognition are powerful motivators in life saving systems
  • Privacy preserving tools like ZK proofs are not just cool ,they’re necessary

What's next for HaemoAI

This is just the beginning. We plan to:

  • Pitch HaemoAI to NGOs and health tech accelerators for early pilot exploration
  • Collaborate with real donors and caregivers to validate our assumptions
  • Work with developers to build a lightweight, scalable MVP in the next 2–3 months
  • Partner with researchers to train region specific LLMs and tune our donor prediction models
  • Advocate for a national blood coordination layer where HaemoAI could plug in

Our vision: to become India’s automated intelligent blood booming infrastructure and inspire other countries to do the same.

Because when blood arrives on time,
hope arrives too.

Built With

  • aes-256+hsm+zkp-security
  • bert-triage
  • e-raktkosh/aadhaar/google-maps-apis
  • e-raktkosh/google-maps/twilio-apis
  • fine-tuned-llama-2-7b-+-lora-(who/icmr-datasets)
  • fine-tuned-llama-2/bert
  • firebase/gcp/polygon
  • firebase/gcp/polygon-zkevm
  • firestore/postgresql/ipfs
  • nft
  • pwa-offline-mode
  • python/js/solidity
  • react/flask/pytorch
  • react/flask/react-native
  • twilio/sms-fallback
  • webrtc-telemedicine
  • with-aes-256/hsm/zkp-security
Share this project:

Updates

Submission history