-
-
Hemo
Inspiration
The need for blood donation is critical, with significant demand but limited supply. Many individuals are unaware of the need, and in certain regions, finding a blood donor is exceptionally challenging. Hemo is designed to raise awareness, encourage blood donation, and provide valuable information to prospective donors. Leveraging AI, Hemo seeks to empower users by predicting blood donation patterns and offering personalized guidance, aiming to save lives through better donor engagement.
What it does
Hemo is an AI-driven chatbot that assists blood donors with essential information and guidance. It allows users to input their blood reports, provides information about their blood type, and suggests compatible types for donation. Hemo also predicts blood donation patterns using data analysis, helping blood centers prepare for potential needs. Furthermore, it helps users find nearby blood donation centers, making it easier for them to contribute to their communities.
How we built it
We developed Hemo using a range of AI and machine learning tools. The chatbot functionality is powered by the Google Gemini Large Language Model for conversational interactions. We used Google Places API for mapping nearby donation centers. For document-based question-answering, we integrated an open-source model from Hugging Face, allowing users to upload and analyze their medical reports. The data analysis aspect of Hemo was built using Python’s Pandas library, enabling blood donation pattern predictions.
Challenges we ran into
One of the main challenges was handling medical document data and enabling effective, context-sensitive question-answering. Ensuring accurate prediction of blood donation patterns based on varied datasets also posed difficulties. Additionally, optimizing the model selection and integrating APIs for streamlined, responsive interactions required careful balancing of resources and performance.
Accomplishments that we're proud of
We are proud to have created a tool that directly contributes to social good, addressing a pressing global issue. Successfully implementing AI for both conversational and predictive functionalities in a user-friendly interface is a significant achievement. Hemo’s capacity to analyze medical reports and predict donation trends provides a unique, impactful experience for users, encouraging more individuals to become blood donors.
What we learned
Through this project, we deepened our understanding of generative AI and large language models in the healthcare domain. We learned to work with medical data responsibly and gained insights into the integration of open-source models and APIs for a seamless user experience. Additionally, we appreciated the importance of balancing model performance with resource efficiency to reduce the environmental impact of AI operations.
What's next for Hemo
The next steps for Hemo include improving its document analysis capabilities by incorporating more specialized medical models. We also aim to refine the blood donation prediction algorithm to enhance accuracy. Expanding the app’s reach to include partnerships with blood donation centers and hospitals, and potentially adding support for additional languages, will further Hemo’s impact and accessibility.
Log in or sign up for Devpost to join the conversation.