CHF Buddy AI

interface

Inspiration

Imagine a world where cardiovascular care is as smart as it is compassionate. Introducing SALUS AI – the Cardiovascular Health Coordinator. Our innovative platform tackles two major challenges: patients struggling with medication adherence and fragmented, incomplete data, and physicians overwhelmed by the "spray and pray" approach to care.

What it does

SALUS AI empowers patients with an intuitive voice AI that asks the right questions, sends timely SMS reminders, and even detects snoring patterns for a holistic view of their health. Our dynamic dashboard brings granular insights—ranging from pain levels to medication interactions—directly into the hands of healthcare providers. With seamless EPIC FHIR integration and plans to incorporate wearable data from top devices like Fitbit, Apple, and Oura, we ensure that every heartbeat is monitored with precision.

Revolutionize cardiovascular care with SALUS AI—where technology meets tailored, proactive healthcare.

How we built it

Informed by personal experience in managing cardiovascular disease and the complexities of remote 24/7 healthcare, patient adherence, and physician insights, we set out to empower patients and their physicians to stay in sync by making healthcare tracking, medication reminders, and question answering seamless and effortless.

The team split tasks as follows:

Naga V Kanaparthy (@impish1) built the support agent on ElevenLabs, iterating on prompting, tools configuration and user experience evaluation. Additionally, Naga prepared documentation edited our demo video, and created graphics and other assets such as presentation materials.
Muntaser Syed (@muntaser836) built the data management system, back-end infrastructure, and wrote Salus AI's REST API
Sergio Leon (@cbnsndwch) created Salus AI's front-end with Lovable.dev, configured GitHub CI/CD, integrated Eleven Labs' React SDK, and implemented client-side data fetching and LLM tools

Challenges we ran into

With a very short time frame to build a functional user experience that put patient's success first, we were faced with building in parallel and with very tight coordination across the team. We needed to rely on low-fidelity coordination tools such as a shared document and hasty verbal communication. Along the way, we encountered challenges in integrating voice AI, client-side LLM tool calling, access control, and overall solution security.

Accomplishments that we're proud of

Through a combination of mature technologies such as Google Cloud, MongoDB Atlas, and React Router, novel cutting-edge AI/ML products such as Lovable.dev, ElevenLabs Generative Voice AI, and our team's extensive experience building digital products, we were able to iterate at incredible speed and solve every challenge we came across.

Our team formed organically during the preparation phase for the event and we are extremely excited about the polished user experience we were able to put together in such a short time.

Finally, and most importantly, we are extremely proud of the fact that the solution we've built during the event is already a qualitative improvement in the life of a person living with cardiovascular disease, and that further improvements to Salus AI will continue to improve their healthcare experience.

What we learned

We take a few incredibly high-valued learnings from this exercise:

Hackathons are fun! There's a very satisfying quality in choosing a real-world problem and hacking away at it until there is a usable, lovable product on the other side of it.
It is easy to underestimate the power that lies in high-pressure, intense, in-person collaboration. An exercise like the one we completed tonight would not have been the same had we taken the challenge purely online.
Developer tools sitting slightly above and simplifying the experience of building with LLMs, voice, and multimodal AI models in general pack a powerful punch. We were able to complete what we set out to do in great part because we had an arsenal of tools at our disposal that saved us many hours of manual work and allowed us to iterate on our solution with mere minutes between major architecture and implementation changes

What's next for Salus AI

We look forward to continuing to improve the user experience, integrating with popular healthcare and fitness tracking technologies, and crossing over from the patient side of the problem space to also deliver efficiencies and qualitative improvements to healthcare providers. With the help of AI technologies and standing on the shoulders of the giants that have come before us we believe Salus AI will make an immense difference for patients with cardiovascular disease and their healthcare providers.