CareCast
I helped Team Storm Logic build a messaging-based climate-health companion. We won first place at the 2025 Health in Climate NYC Hackathon at Cornell Tech.
The product
CareCast combines local weather conditions with personal health context to send relevant alerts during extreme heat. The prototype delivers guidance through messaging rather than requiring a new app.
My contribution
I contributed software engineering and MLOps alongside teammates working in medicine, climate science, planning, and product. The implementation uses Node.js, Twilio, Redis, BullMQ, and Gemini to connect weather data, conversational responses, and scheduled messages.
From weather data to messages
The prototype uses an Express server for API requests and incoming messaging webhooks. Weather data comes from the National Weather Service, while Twilio connects the application to WhatsApp.
BullMQ and Redis separate scheduled work from the web server. Background workers process weather checks, recurring check-ins, and outgoing messages, including retry handling and delivery tracking. PostgreSQL with Prisma stores the application’s structured data.
The team
Team Storm Logic included Sandya Venugopal, Shivani Kadakia, Maya Chung, Tiffany Luo, Benedict Ho, and Muhammad Ahsan.