MediMatch

Inspiration

MediMatch was inspired by a personal story—my dad's journey through a misdiagnosed eye condition. Originally told it was caused by stress, he took it upon himself to research further and eventually discovered the real culprit: H. pylori bacteria. His persistence and self-guided research ultimately led to a proper diagnosis and cure. Our project aims to make this type of critical investigation more accessible. Medi-match empowers users to explore alternative explanations and treatments through AI, reducing the effort and confusion that my dad had to endure.

What it does

MediMatch is an AI-powered tool that helps users explore potential medical explanations beyond a surface-level diagnosis. By inputting symptoms or concerns, users receive curated suggestions backed by medical literature and AI-generated analysis. It’s designed to assist—not replace—medical advice, giving users an extra layer of information and prompting more informed conversations with healthcare professionals.

How we built it

We built MediMatch using a combination of large language models, frontend UI frameworks, and medical data APIs. One of our core tools was GPT for natural language understanding and generation. The development process involved rapid prototyping—starting with brainstorming over 20 different ideas with the help of AI, followed by vibe-coding and down-selection. The team worked on separate forks and merged our code using Git to maintain project organization.

Challenges we ran into

One of the major hurdles we faced was coordinating version control using Git. Initially, we had conflicts and confusion when pushing to the main branch. We resolved this by setting up individual forks and later merging them, which kept our codebase cleaner and more manageable. Another challenge was integrating AI responses in a medically responsible way, which we’re still improving.

Accomplishments that we're proud of

We’re proud of successfully creating a functional AI tool that mimics the critical thinking process my dad used. From developing a clean UI to integrating intelligent AI outputs, we’ve laid a strong foundation. The brainstorming phase, AI-assisted idea generation, and the actual implementation gave us confidence that we're onto something meaningful and scalable.

What we learned

One of the most valuable lessons we learned was how effective AI can be in early-stage product ideation. By using GPT to generate 20+ ideas and then vibe-coding them as a group, we cut down on decision time and prototype iterations. We also learned how important Git workflows are for collaboration and how structured brainstorming can make a massive difference in team productivity.

What's next for MediMatch

Next, we aim to fully integrate AI into every stage of the diagnostic support process, ensuring it's helpful without being misleading. We plan to squash existing bugs, refine our user interface, and streamline the user experience. Our ultimate goal is to make Medi-match a seamless, reliable assistant that can truly support patients and caregivers in navigating complex health information.