Seconds - Giving Slightly Imperfect Food a Second Chance

Seconds rescues “imperfect” food from waste. Our platform connects suppliers to buyers, using AI models to assess quality, price, and value—making food affordable and sustainable.

Comment

Teammates - Kenneth Choi, Ryan Lung, Matthew Bevins, Agastya Bassi

Inspiration

Walking through grocery aisles, we noticed how much edible food was thrown away simply for being a little “ugly.” At the same time, communities face food insecurity. Seconds bridges this gap — using AI to identify, list, and redistribute perfectly good food.

What it does

Seconds is a two-part platform with a mobile app and a web dashboard:

🧾 For suppliers

  • Snap a photo of your food item.
  • Our AI model determines edibility, approximate weight, nutritional value, and expiration date.
  • Instantly list the item on the Seconds marketplace.

🛒 For buyers

  • Browse discounted imperfect foods on the app.
  • Find affordable and fresh food from nearby suppliers.
  • Purchase directly through the platform.
  • Enjoy fun facts and helpful info about their food, generated using the Gemini API.

📊 For analytics

  • Suppliers can view total revenue, sales history, and real-time inventory insights through the web dashboard.

How we built it

AI/ML:

  • Trained a custom convolutional neural network (EfficientNet) to detect blemishes and spoilage in food.
  • Hosted the model on a FastAPI server.
  • Used Gemini and LLM APIs to estimate weight, nutritional value, and expiration date.

🧠 AI Text & Engagement:

  • Integrated the Gemini API to dynamically generate fun facts, storage tips, and contextual info for each food item — making the experience both informative and engaging.

📲 Mobile App (Suppliers & Buyers)

  • Built using Flutter & Dart.
  • Integrated Firebase Auth for secure login and Firebase Storage for image uploads.
  • Real-time listings powered by Firestore.

💻 Web App (Suppliers)

  • Dashboard built with React and TypeScript.
  • Displays sales analytics and inventory metrics in real time.
  • Pairs with the mobile app to allow businesses to manage their inventory + listings.

Challenges we ran into

Coordinating the database schema was difficult especially because we were working on both the web/mobile app at the same time. We spent a lot of time training an accurate model via food image data. Additionally, we spent a good amount of time fixing CORS errors when we hosted our custom CNN via FastAPI.

Accomplishments that we're proud of

We are proud to have been able to build a complete platform that uses AI at multiple angles. We were able to build our own CNN as well as use Gemini for text generation and some minor image processing.

What we learned

We learned how to simultaneously develop for both web and mobile at the same time. It was a very rewarding experience and allows for a more complete platform for our end users.

What's next for Seconds - Giving Slightly Imperfect Food a Second Chance

  • Integrate geolocation to show nearby deals.
  • Expand AI capabilities to more food categories.
  • Partner with nonprofits and food banks.
  • Add automated pickup and delivery options.
  • Use Gemini to power chatbot-based assistance for suppliers and consumers.
  • Allow for scale with potential monetization.

Built With

Share this project:

Updates

posted an update —

Great experience and each person brought so much to the team. Loved the UI design and database security on the app and website, it looks so sleek and functions smoothly. Glad that I finally got the AI training and API to work. Great demo video as well, was a lot of fun filming and getting the video together. Had a blast working with you guys! RedBull came in clutch

Log in or sign up for Devpost to join the conversation.