← Back to selected workAI · Full-Stack · Ongoing · Currently building

Wingtern

An AI-powered student growth companion focused on helping students learn, plan and grow through personalized guidance and intelligent workflows.

StackNext.js · TypeScript · Supabase · OpenAI
Evidence88 / 100 evaluation · OpenAI × Outskill AI Builders Hackathon
Wingtern AI-powered student growth companion dashboardWingtern / supplied media
01 / Problem

Start with the reason to build.

Students often have goals, questions and plans spread across disconnected tools. Wingtern explores how an intelligent companion can make that growth journey more coherent.

02 / Context

What shaped the work.

Wingtern is an ongoing personal product and was evaluated at the OpenAI × Outskill AI Builders Hackathon.

03 / Solution

A system, not a feature demo.

The product brings together an approachable frontend, authenticated user flows, a dashboard and OpenAI-powered guidance so students can learn, plan and grow inside one evolving system.

04 / My contribution

Be precise about ownership.

I built the full-stack MVP: the Next.js frontend, Supabase authentication and database, OpenAI integration, dashboard, onboarding and end-to-end integration.

05 / Implementation

The parts that made it real.

  • Next.js and TypeScript application structure
  • Supabase for authentication and persisted application data
  • OpenAI integration for personalized guidance
  • Dashboard and onboarding flows for the end-to-end product experience

Architecture is represented here as a focused system sketch: the important interfaces and decisions are more useful than a decorative diagram.

06 / Challenges

Where engineering judgment mattered.

  • Turning an open-ended product idea into a focused MVP
  • Connecting AI assistance to real user flows instead of treating it as a standalone demo
  • Balancing ongoing product exploration with a maintainable full-stack foundation
07 / Tradeoffs

Every system carries a choice.

The MVP favors a small, integrated stack that keeps product iteration fast: Next.js for the application surface, Supabase for the core backend primitives and OpenAI for the intelligence layer.

08 / Outcome + learning

88 / 100 evaluation at the OpenAI × Outskill AI Builders Hackathon. The product is still being developed.

  • AI features are strongest when they are attached to a clear workflow.
  • A product foundation should leave room for learning from use, not over-specify the first version.
  • Onboarding is part of the product system, not just a first screen.
Back to selected workView project on GitHub ↗