About the Project A fully-featured nutrition mobile app combining AI-driven meal planning with calorie tracking, progress charts, smart notifications, challenges, and personalized reports. The AI analyzed each user's activity, habits, goals, and food preferences to continuously adjust meal plans across a library of 3,000+ meals.

My Role Solo founder, product designer, and product owner. I designed, launched, and scaled the product over 2 years — from zero to 5,000 MAU, managing both product direction and end-to-end design.

I developed a soft, calming visual system that reflects the idea of mindful, everyday wellness. The interface is built around warm neutrals and natural green tones, creating a sense of balance and trust. Typography is simple and highly legible, supporting quick interactions and effortless scanning of nutritional data. Rounded UI elements and gentle spacing reinforce a feeling of comfort and accessibility, making the experience feel supportive rather than overwhelming. The result is a cohesive brand identity that feels personal, lightweight, and encouraging—positioning the product not just as a tracking tool, but as a daily companion for building healthier habits.

The Problem Most nutrition apps offer generic plans that don't adapt to who you actually are. AI Nutrition was built around one core idea: your diet should be as individual as your lifestyle. Goals, activity level, food preferences, daily habits — all of it feeding a plan that continuously adjusts to you.

Meal logging that people actually use. Logging is the most repeated action in any nutrition app — and the most abandoned. I placed it in a persistent quick-action menu (a central + button in the nav bar) giving users instant access to all four input methods: by product, recipe, text description, or photo. All methods lived on the same screen, each reinforced with an icon for fast scanning. The photo flow split into two clear sub-options — scan a dish or scan packaging — reducing errors before they happened. If users couldn't photograph their meal, they could simply describe it and AI would calculate the macros automatically.

Search across 10,000+ items without overwhelm. Rather than a flat database, search was personalized from day one — prioritizing groceries aligned with the user's goal, and surfacing previously logged items for faster re-entry. The more you used it, the smarter it got.

An AI coach that felt native, not bolted on. The health coach had full access to user data — logs, goals, activity, trends — so it could make genuinely informed recommendations: diet adjustments, grocery swaps, calorie reminders. It complemented every other widget in the app rather than existing as a separate chat feature.

Onboarding 20 questions across 4 groups — but designed to feel like momentum, not a form. Between questions I showed users projected results at 7 and 30 days, how many recipes matched their food preferences, statistics from other users, and a preview of key features. The goal was to make personalization feel tangible before they finished setup. Completion rate: 73%.

Recipes The recipe constructor — where users could build their own nutrition plan by combining app recipes, AI suggestions, and their own ideas. It's where the whole system came together: the content library, the intelligence, and the user's own agency over their health.