Saarthi AI

"Predicting hope. Delivering Life."

Comment

Inspiration

I started with one question: What if we could replace a patient’s fear?

In India thousands of Thalassemia patients depend on regular blood transfusions to survive. Yet, it is always very difficult to find a compatible donor which makes this entire process quite stressful, unpredictable, and emotionally exhausting for families.

We were inspired by the incredible efforts of groups like Blood Warriors and the quiet resilience of the families behind every patient. But we also saw how broken and overwhelmed the system is—scattered records, delayed responses, and missed opportunities to save lives.

That’s when the idea of Saarthi AI was born. The name “Saarthi” comes from Sanskrit which means a charioteer. A charioteer is the one who guides a warrior through battle. While we can’t cure Thalassemia, we know we can guide and support the people who are fighting it.

So we imagined something more than just a tool. We envisioned a companion—one that’s always one step ahead, helping families and caregivers to be prepared before the crisis hits.

What Saarthi AI Does

Saarthi AI is not just another app—it’s a smart, supportive system designed to make life easier for patients, donors, and care organizations.

Smart Donor Matching

To predict who is likely to be available and when, we use past donation patterns, donor location, and other key data. This helps us to match patients with the right donors quickly as it ==cuts off the waiting time from days to just a few hours==.

Support for Patients, Anytime

Asha Bot, our friendly chatbot, is always there for patients. Whether it is for helping with daily care tips, answering questions, or reminding them about their transfusions. It is like you have someone to talk to you 24/7.

Helping Organizations Work Smarter

Groups like Blood Warriors get a single dashboard to track patient needs, organize donation drives, and keep an eye on blood availability in real time—all in one place.

Keeping Donors Involved

We show donors the real impact of their actions. It shows them the number of lives they’ve helped in saving. It motivates them to come back with badges, appreciation messages, and personalized updates.

How we built Saarthi AI

I built Saarthi AI on a solid and secure tech stack so it's easy for people to use and holds up under real-world conditions.

  • Frontend: I went with React.js to build a clean, mobile-first web app that works on any device right from the browser, so there's no need to mess around with app store downloads.

  • Backend: I used Node.js with Express.js to handle things like real-time messages and notifications, with Firebase helping out on the push notifications side. This makes chatting and updates smooth and quick.

  • Database: For storing user info, I picked MongoDB because it's flexible, and added Firestore on top to handle live logs — things like chats or quick status updates.

  • The AI Brain: At the heart of Saarthi is what I call the Saarthi Suitability Score (\(S^3\)). I wrote it in Python using Scikit-learn, then shared it through a Flask API. This score helps rank donors based on three main things:

    • Donor Profile (\(P_d\)): their donation history and eligibility
    • Location Proximity (\(L_p\)): how close they are to the patient or hospital
    • Context (\(C_t\)): stuff like traffic, weather, or their usual routine

The formula looks like this, weighted for fairness: $$ S^3 = w_1 P_d + w_2 L_p + w_3 C_t $$

  • Integrations: I also hooked it up to Google Maps so we can figure out donor routes in real-time, and to WhatsApp Business for sending critical alerts fast. ## Challenges we ran into Every meaningful journey has hurdles on its way, and ours was no different. Some of the biggest challenges that we tackled are:

  1. Starting Without Data Our AI system needed data to work—but at the beginning, we had none. So instead of waiting, we built a smart, rule-based system that could function right away. Over time, this system will keep on collecting data which will help our AI to train and become ==even better==.

  2. Reaching Users in Smaller Cities We knew that many of our users would come from Tier-2 and Tier-3 cities, where using apps isn’t always easy. Which is why we designed an interface that would be super simple to use. We added support for multiple languages and have planned to include SMS and IVR options for people who don’t use smartphones.

  3. Protecting Sensitive Information We’re dealing with personal medical data, and we took that responsibility very seriously. We spent a lot of time making sure everything is secure—end-to-end encrypted and built with privacy at its core, following India’s data protection laws.

Accomplishments that we're proud of

There are a few things that make us especially proud:

  1. We built something that prepares ahead of time instead of reacting at the last minute. That’s a big shift in how blood donation systems usually work.

  2. We created the Saarthi Suitability Score (\(S^3\)) — a formula that helps match donors more accurately — and we’ve already tested it with a working prototype.

  3. We didn’t just focus on one group. Our solution brings together patients, donors, and support teams into ==one shared platform==.

And above all, we stayed true to the spirit of Saarthi — being a guide and companion for those in need. That idea shaped every decision we made.

What we learned

This project taught us a lot, but here are the biggest lessons:

  • Empathy matters. We didn’t set out to just build an AI system—we wanted to ease people’s fears. That goal kept us grounded throughout.

  • A strong identity makes a big difference. The name Saarthi gave our project meaning and helped us stay focused on who we’re building it for.

Sometimes, the best solutions are the simplest ones, and especially it becomes even more simple when you’re ==designing it for the people who are already going through so much==.

What's next for Saarthi AI

Until now, we’ve only taken the first step. Here’s where we are planning to go next:

  1. Pilot Launch We’re partnering with Blood Warriors to test Saarthi AI in one city. This will help us see how it performs in real situations and start collecting ==real-time data==.

  2. Smarter AI Rollout Once we have enough data, we’ll train and improve our prediction model to make donor matching even more accurate and faster.

  3. Connecting Nationally We aim to link with India’s national platforms like e-RaktKosh so we can help build a stronger, more connected blood donation system.

  4. Expanding Support Looking ahead, we want Saarthi to do more—like offer mental health support, connect patients with each other, and provide trustworthy medical info. We want it to be a true companion for every Thalassemia warrior.

Built With

Share this project:

Updates

Submission history