Multimodal OTC medication safety agent
As an international student in the US, I experienced firsthand that buying over-the-counter medication without understanding drug interactions is genuinely dangerous. Tens of thousands of people end up in emergency rooms each year from chemical interactions between common OTC drugs. This wasn't just my problem. Everyone with limited healthcare access faces it.
We built a multimodal AI agent (voice, image, video) that checks a user's existing medications, allergies, and dietary habits before recommending OTC drugs. The agent prevents dangerous interactions and sends dosage reminders. I assembled the founding team in the US, recruiting developers, a designer, and a strategist directly.
This was my first time building a team from scratch in a new country, in a new language, with no existing network. The product worked. Users loved it. But the economics didn't. Running multimodal AI (voice + image + video) at the quality level healthcare requires was too expensive for a consumer app without a clear revenue model. I learned that product-market fit isn't enough. You also need product-economics fit.