Team MoonRock
An AI use case for SafeSpec: turning food scans from a near-infrared spectrometer into predictive allergen risk models for manufacturers and restaurants.
The brief
Sunstone Management's AI hackathon asked teams to develop predictive modeling and data intelligence solutions for a real startup. Ours was SafeSpec, whose NIR spectrometer scans food for allergens.
Each scan answers one question for one person. Our pitch was to treat every scan as a data point: pool them with metadata like ingredients, preparation methods, and handling environments, and learn why allergens show up where they do. That turns a consumer device into an industry data product.
Four AI opportunities
Predictive allergen risk modeling
Use scan data to predict cross-contamination, pinpoint high-risk sources, and stop allergic reactions before they happen.
Intelligent food profiling
Map hidden allergen patterns across ingredients, prep methods, and NIR readings, learning how foods vary to build an evolving safety database.
Industry-powered learning network
Aggregate anonymized industry data to sharpen detection accuracy and cut down on blanket “may contain” labels.
Personalized safety
Adapt to each user's allergy profile with personal alerts, safe menu suggestions, and predictive warnings wherever they eat.
How it would roll out
A five-step loop. The model keeps improving as more people scan.
- CollectAggregate user allergen results and scan metadata.
- TrainStore it and train a model linking allergens, foods, locations, and manufacturing processes.
- Find riskSurface the high-risk points in supply chains and kitchens.
- ImproveTrain staff and adjust supply chains to prevent cross-contamination.
- ReinforceKeep collecting scans and retrain on newer data for accuracy.
↺ step 05 feeds back into step 01
Who it helps
Manufacturers
- See real allergen contamination trends by region.
- Find weak points in the supply chain.
- Reformulate products to remove accidental allergens.
- Earn consumer trust with a second, independent check on ingredients.
Restaurants
- Prevent cross-contamination by knowing exactly where the risk is.
- Raise customer satisfaction and reputation for safety.
- Grow revenue as higher safety standards bring in allergy-conscious diners.
- Market that safety with data behind it.
More accurate labels, safer food systems, and new markets for allergy-safe products. A win for consumers and for industry.
The team
- Vyom Shah
- Sawyer Valin
- Seth Flucas
- Kent Kawashima
- Rachit Yadav
- Akarsh Ranjan
- Varshini Vijay
- Marisa Chacon