Problem
Pick two food cultures and the app blends them into a fusion recipe, drawing on 28 cuisines.
Built with a team of five, Aun Raza, Ibraheem Iqbal, Anas Bataweel, Ismail Shafie and me, at GreatUniHack 2025. For the result to be usable it has to be structured, not free-form text: recipes with ingredient-origin tracking, plus generated flag imagery for each pairing.
The goal, straight from the project brief we wrote ourselves: bridge cultural divides through food, using AI to generate fusion recipes across cuisines rather than just listing them side by side.
Decisions
Next.js 15 and TypeScript end to end. Google Gemini does the generation, returning structured recipes with ingredient-origin tracking rather than plain text.
Engineered properly despite the hackathon time box: API caching with TTLs, rate limiting, and error boundaries, plus full keyboard and screen-reader accessibility.
We split the work based on expertise: myself and Aun worked on the front end, Ibraheem worked on the back end, Anas worked on integrations and the API, and Shafie was the team's presenter.
There's no live deployment: the repo is the only link. Deployment to Vercel was planned but the clock ran out before submission.
What I'd do differently
We'd spend more time on design thinking, properly fleshing out the pre-build activities before jumping into code. Future work would add more social features: sharing, recipe ratings and community recipe submissions, which would mean adding user authentication and accounts.