DiagnosAI

Inspiration

Our group was deeply inspired by the powerful potential of AI in addressing global healthcare challenges, specifically focusing on changing the accessibility issues that often impact people. The idea for DiagnosAI emerged from our passion for leveraging technology to make healthcare more reachable, especially so that anyone can use it. The Sustainable Development Goals (SDGs) underscore the need for equitable healthcare, and we envisioned DiagnosAI as a tool that could bridge gaps in access to medical knowledge and support.

What it does

DiagnosAI is an innovative AI-driven chat platform designed to provide preliminary medical consultations and health-related information. It utilizes advanced natural language processing to understand user inquiries and offers insights based on medical data. This tool is particularly beneficial for individuals in remote or underserved areas, where access to medical professionals is limited, as anyone can access it and get some sort of medical health.

How we built it

We integrated various AI technologies, including a multinomial logistic regression model and natural language processing plus machine learning algorithms, to create a user-friendly and efficient system. The core of DiagnosAI is built on top of Google's AI's Vertex AI for Gemini Pro, ensuring reliability and scalability. The frontend is built with React, while the backend is built node.js.

Challenges we ran into

One of our biggest challenges was navigating through the frontend API provided by chatengine.io, specifically in linking specific chats to automated responses. The chatengine.io API ended up being quite limited in its various props and automated message sending, forcing us to think outside of the box. We also faced technical difficulties in integrating different AI components into a cohesive system, integrating the model we trained to work with the gemini api as well as the rest of the front end.

Accomplishments that we're proud of

We're proud of creating a tool that has the potential to democratize access to healthcare information. Our successful integration of complex AI technologies to deliver solutions in healthcare is something we very proud of. Having also figured out a weird fetch request to the remote API, we consider the full stack application that we built a success.

What we learned

Through this project, we gained valuable insights into the intricacies of healthcare technology as well as AI technology, and incorporating LLM's such as gemini-pro. We also learned a great deal about incorporating node.js as a backend with a model, and tuning a gemini model in node.js

What's next for DiagnosAI

Looking ahead, we aim to expand DiagnosAI's capabilities to include personalized health tracking and integration with human-to-human contact, connecting doctors and patients. We also plan to enhance our model with more diverse medical conditions and treatments, focusing on rare diseases that are often overlooked.

Built With

• ai/ml
• css
• express.js
• google-cloud
• html
• javascript
• node.js
• python
• react
• vertex-ai-api