Swasth Connect- “RAG Doctor “

• 

  RAG Doctor

Inspiration Adapt and change is where most amount of value can be created, so we want to build a user friendly tele medicine app that not only focus on quality, but also on privacy and accessibility i rural areas where there is no network or less network. With millions of people living in rural and underserved areas with limited access to quality medical services, we wanted to create a solution that bridges this gap. Our goal was to develop a system that not only leverages advanced AI technologies but is also simple enough for anyone to use, regardless of their technical knowledge. We envisioned a platform that brings real-time, value-based healthcare to your fingertips, eliminating barriers like distance, cost, and complexity.

What it Does SwasthConnect - "RAG Doctor" is an AI-driven telemedicine platform that provides instant, real-time symptom analysis and personalized healthcare advice. Using Retrieval-Augmented Generation (RAG), the system retrieves relevant medical data and combines it with AI-enhanced diagnostics to offer accurate, context-specific responses. It’s designed to cater to a wide range of users, from rural populations to women, children, and individuals managing chronic conditions. The app works offline on your mobile device, ensuring privacy, low data usage, and instant access to healthcare services without requiring a medical license or technical expertise.

How We Built It We built SwasthConnect - "RAG Doctor" using a combination of advanced AI technologies and user-centric design principles. The core system is powered by Local LLaMA3.1, running on a local Nvidia GTX1070, which ensures quick processing and privacy. We utilized a Vector Database to store and retrieve medical and symptom data, allowing the system to provide highly contextual and accurate responses. The platform integrates with a Streamlit UI for a seamless user experience and uses a FAST API Server to handle requests and responses. Additionally, we employed SQLite3 for secure authentication and storage of medical records. This robust yet lightweight architecture makes it easy to deploy and use on any mobile device.

Challenges We Ran Into Running advanced AI models locally on a device without compromising performance was a technical hurdle we had to overcome. We also needed to ensure that the user interface was intuitive enough for non-technical users, including those in rural areas with limited digital literacy. Additionally, integrating a comprehensive and secure database of medical history and symptoms required meticulous attention to data management and security protocols.

Accomplishments That We're Proud Of We are proud to Enjoy the whole Hackathon process and working together on a Problem in limited tie constraint, building a developed system that brings advanced, AI-driven healthcare within reach for everyone and Creating a solution that operates efficiently on local hardware, ensuring privacy and low data usage, is a significant accomplishment. We successfully designed an intuitive user interface that caters to a diverse user base, including those with no technical knowledge.

What We Learned Through this project, we learned the importance of balancing advanced technology with user-centric design. Developing a system that combines AI and RAG in a way that is both powerful and easy to use was a significant learning experience. Our experience reinforced the value of integrating privacy and efficiency in AI systems, especially when dealing with sensitive medical data.

What's Next for SwasthConnect - "RAG Doctor" Our next step is to expand its capabilities and reach. We plan to incorporate more comprehensive medical databases and enhance the AI model for an even broader range of diagnostic capabilities. We aim to integrate additional features like multilingual support to cater to non-English speaking users and expand the platform’s reach globally. Ultimately, our goal is to make SwasthConnect a ubiquitous tool in global healthcare, ensuring everyone has access to quality, real-time medical advice regardless of location or socioeconomic status.

Try it out GitHub Repohttps://github.com/jayaramakarthikeya/RAG-Doc

Built With

• api
• fastapi
• langchain
• llm
• python
• sql
• sqlalchemy
• sqlite
• streamlit