NutriScan AI
A student-focused nutrition prototype, co-built at HackWestern 12. It explores photo-based AI input and budget-friendly meal planning in one web app.
A healthier plate.PHOTO → INSIGHT → MEAL PLAN
The starting point
The team set out to make nutrition guidance easier for students to explore, with meal suggestions that account for a limited budget. The time constraint: build and present the idea in 36 hours.
What I built
I co-developed the React and TypeScript interface and helped integrate multimodal AI into the experience. The project uses Google Gemini 2.5 Flash and DigitalOcean Gradient AI, with Tailwind CSS for the interface.
Under the hood
The prototype takes photo input and generates personalized meal suggestions. It is a hackathon exploration, not a validated diagnostic tool; the useful engineering challenge was connecting image input, model responses, and a readable result.
What stayed with me
NutriScan won the Sun Life Best Health Hack and Best Use of DigitalOcean AI awards at HackWestern 12. The sprint pushed the team to coordinate interface work and AI integration around a working demo.