WASTEWISE
The Project
One day I noticed my mom cutting off corners of the bread crust when she was making a sandwich. It seemed like unnecessary wastage. I did a survey and sent findings to some bread companies asking them to remove the crusts, top and bottom pieces of the bread loaves at the factory itself and repurpose or recycle them. WasteWise idea came much later. I guess, because by then I had started consciously looking at food waste around me that could be avoided. In my own school I noticed bins filled with food at the end of meals. I started interacting more with the kitchen staff to explore if the over preparation or kitchen over ordered raw material part of the waste could be reduced. I realised that their planning relied on experience and intuition, and when things went wrong, there was no data to explain why. The waste simply disappeared into a bin, unnoticed and unmeasured. It was similar for grocery stores and supermarkets. That to me was a problem arising out of insufficient decision making data since it was totally subjective and didn’t let the operators know what was causing the wastage. This motivated me to build a solution that does not blame people or add pressure to already busy kitchens or grocery stores busy with their business operations but quietly supports them without changing their workflows and working quietly in the background. Food waste is a $1 trillion global problem. Restaurants/supermarkets generate over 400 million tonnes annually. WasteWise is a passive AI system that detects inefficiencies, waste losses and food degradation across kitchens, storage, and logistics before waste occurs. Waste monitoring industry presently competes on higher precision, manual weighing and reporting sophistication with substantial workflow disruption. We eliminate manual weighing and workflow disruption. We reduce hardware/ SaaS cost drastically with deployment cost at USD 100 and reduce staff dependency. We raise real-time visibility across intermediate food prep which is where 60-70% wastage occurs and existing systems don’t cater to. We shift from reactive waste measurement to hands-free, cross-stage, predictive loss prevention without workflow disruption. During initial testing and deployment stage, our Image recognition accuracy exceeded 98%. Quantity estimation accuracy exceeded 90%. The system is expected to lead to a waste reduction of more than 30%. Version 1 is tested and deployed in three kitchens. Version 2 expands into predictive degradation monitoring for food storage/logistics and integrate sensor data like temp/humidity. It will trigger dynamic discounting alerts, suggest procurement adjustments, and incorporate cultural demand spikes. For NGOs and shelter kitchens, the system will always be provided at hardware cost only.
