Back to Engineering Projects
Personal Health🤖 AI Applications
AI Diet Planner
An AI-powered nutrition assistant that generates personalized meal plans, calorie recommendations, and nutrition insights based on user preferences and goals.
Benchmark Results & Metrics
Meal Options
10k+ Recipes
Macro Precision
95%
Daily User Growth
+12%
Query Latency
~1.2s
01 // Overview
AI Diet Planner is a wellness companion app designed to construct custom dietary schedules. It adapts in real time to macro-nutrient targets, food allergies, and local ingredient availability.
02 // The Problem
Standard calorie calculators ignore food preferences and cultural restrictions, leading to low user compliance and unsustainable diets.
03 // System Architecture
Caloric & Macro-nutrient Solver Architecture
The app collects health profile parameters on-device and leverages Gemini APIs to draft dynamic, balanced meal lists that strictly adhere to nutritional requirements.
ARCHITECTURAL DATA FLOW:User Metrics -> Basal Metabolic Rate Solver -> Gemini Macro Builder -> Recipe Database -> Mobile App
Architecture Components:
React Native Health Dashboard
Calorie Requirement Solver
Gemini Recipe Generator
Macro Tracker Database
04 // Implementation
Created using Expo for rapid cross-platform deployment. Leverages client-side storage for tracking daily macro consumption and remote LLM queries to fetch tailored recipes.
05 // Execution Workflow
11. User sets health profile and diet goals.
22. Basal metabolic rate and macro split calculated.
33. Gemini generates weekly meal configurations.
44. User logs food consumption throughout the day.
55. AI dynamically adjusts remaining daily meal targets.
06 // Technology Stack
Frontend & App
React NativeExpoTypeScript
AI Orchestration
Gemini AIClient Storage API
07 // Technical Challenges
⚠️Ensuring meal generation remains safe and aligns with dietary restrictions like celiac disease or nut allergies.
⚠️Optimizing mobile bundle sizes for fast downloads.
Lessons Learned
- •Strict system prompts are critical to prevent LLMs from suggesting hazardous ingredients for allergy profiles.
- •Caching recipe steps prevents repetitive API overhead.
Future Improvements
- •Integrating barcode scanning for automated grocery item logging.
- •Syncing with Apple HealthKit and Google Fit for dynamic energy expenditure tracking.