Food labels are difficult to interpret, while generic nutrition advice ignores the fact that the same product can mean different things for different people.
A personalized food-label interpreter that changes the health read based on the person, not just the product.
Food labels are difficult to interpret, while generic nutrition advice ignores the fact that the same product can mean different things for different people.
I made user context the first-class input, then grounded the analysis in structured nutrition data instead of producing a generic health score from the label alone.
The product became a live Custom GPT and provided the product-learning foundation for the later Medicine Helper build.
Personalization becomes valuable when it materially changes the decision, not when it merely changes the wording of the answer.
Open the full architecture, modeling, validation and lessons.