How NTU FRESH is using AWS to build predictive food safety at scale

This article describes how NTU FRESH at Nanyang Technological University is leveraging AWS services to build a predictive food safety platform that replaces reactive testing with real-time, data-driven assessments.

• IoT sensors capture volatile organic compounds, pH, and water activity readings transmitted securely via AWS IoT Core
• AWS Glue orchestrates ETL of laboratory and omics data for feature engineering and model preparation
• Amazon SageMaker AI trains predictive models correlating sensor signals with food safety outcomes and deploys inference endpoints
• Dynamic shelf-life modeling predicts precise safety windows, reducing food waste by 14.8% in pilot programs
• Amazon Bedrock synthesizes model outputs into plain-English operator guidance for frontline staff
• AWS Amplify dashboard with Amazon Cognito provides real-time shelf-life predictions and risk scores to authorized users
• System architecture is modular and extensible for new organisms, sensors, and food categories without infrastructure changes

The platform transforms weeks of lab work into near real-time intelligence, addressing the one-third of global food production lost or wasted annually.