Mosaic
Mosaic is a values-first connection app built around one idea: meet slowly, connect deeply. Instead of rushing users into swiping, Mosaic helps people build a profile, reflect on what matters, and use AI-guided insights to understand their connection patterns before moving toward future matching.
Client
Independent Concept
Type
AI Product Design · Full-Stack UX Engineering
Year
2026
Role
AI Product Designer & UX Engineer
Process
Reflection before matching.
I started with a simple question: what if a connection app did not begin with swiping?
Most dating apps are optimized for speed: quick profiles, quick judgments, and fast matches. For Mosaic, I wanted to explore a slower path where users could understand themselves before deciding how they want to connect with someone else.
I planned the experience around three core moments: building a values-first profile, discovering people intentionally, and using private reflections to notice themes, values, and patterns over time.
System
A real product, not just a prototype.
I designed and built Mosaic end to end with React, TypeScript, Supabase, OpenAI, and Vercel.
The app supports secure authentication, profile creation, photo upload, private reflections, persistent Save/Pass decisions, user-owned challenges, AI-guided insights, responsive layouts, and live production deployment.
Because Mosaic handles personal reflection content, privacy shaped the architecture. Supabase Auth and Row Level Security protect user-owned data, while OpenAI requests run through a Supabase Edge Function so the API key stays out of the browser.
Product Demo
AI as a reflection aid, not a decision-maker.
The product has two AI moments with clear jobs.
Sam guides users through short reflection questions when they need help putting thoughts into words. Connection Lens works after a reflection is saved, turning it into a simple theme and key values.
I kept the AI intentionally supportive. It does not make decisions for the user, offer therapy, or choose matches. It helps users clarify what they are noticing so they can move toward connection with more self-awareness.
Outcome
A slower way to design for connection.
Mosaic became a live, full-stack AI product that shows both product thinking and technical execution.
The final app supports real accounts, private user data, profile photos, persistent discovery decisions, reflection CRUD, challenge CRUD, AI-guided insights, responsive layouts, and production deployment.
The strongest outcome is the product point of view: connection products do not have to begin with speed. They can begin with self-awareness.
