Health tech · AI
Nutrition Tracker AI
Personal calorie and protein goals, with meals logged by text, voice or chatbot. The numbers had to be right, whichever way they came in.
- Web
- AI analysis
- Calculations
The product
Nutrition Tracker AI helps people reach their nutrition targets. It calculates daily calorie and protein goals from each user’s age, weight and activity level, and gives an AI-assisted breakdown of the meals they log.
The aim was scientifically accurate, personalised guidance that is simple to use day to day.
What had to be tested
- BMR-based goal calculations across different ages, weights and activity levels
- Meal logging by manual entry, voice command and chatbot
- The AI’s nutritional breakdown of each logged meal
- Real-time dashboards showing progress against daily targets
- Onboarding and account set-up
How testing fitted in
The product was built as an MVP in agile sprints, with continuous testing and stakeholder feedback in every iteration.
Goal calculations and multi-modal logging were refined sprint by sprint as testing found issues.
The outcome
What the project reported:
- Fewer errors in dietary tracking through AI-assisted data entry
- Measurable progress for users on weight and fitness goals
- Higher daily interaction through a simpler interface
- Ready to integrate with wearables and meal-planning tools
What we would test today
A calorie goal is a calculation; a meal breakdown is a model’s answer. Today we would test them differently: the calculation against fixed expected values, and the AI’s breakdown against reference answers, including what happens when it invents a food or misses one.
Tools and technologies
- ExpressJS
- HTML
- CSS
- Git
- Figma
- Miro
