For Zou, the idea for dealing with food waste crystallized after he saw mounds of food regularly tossed in the trash in Georgia Tech dining halls. Quickly, the team — including computer engineering majors Abhipsa Ujwal and Nathanael Koh and electrical engineering major Phuc Truong — decided the solution wasn’t so much about dealing with the trash as it was changing behavior.

When users step up to the Raccoon Eyes trash can system and open the lid, the system scans their plate and identifies the volume of food being discarded. Raccoon Eyes then announces the carbon impact of the wasted food and asks the user a specific question, like, “How can we improve the pizza?”

The answers are recorded for review by dining hall staff, offering substantive and direct feedback. Data on the quantity of discarded food feeds a dashboard to help chefs make informed decisions about how much food the staff prepares.

“Early in the semester, we were the talking trash can. We would stand in dining halls and ask, ‘Hey, how was the food? Did you like it?’” Ujwal said. “Some students would talk to me for five or 10 minutes. Yet the text response rate — which is what Georgia Tech uses right now for feedback — is less than 1%. So you can see there’s information there and a desire to share it, but it’s not relating with the current system.”