Dyne

Inspiration

Choosing where to eat is a problem people face almost every day. While there are endless restaurant options, the decision itself often becomes slow and frustrating, especially when multiple people are involved. We noticed that the issue isn’t discovery, but decision friction. Dyne was created to remove that friction and make choosing feel effortless.

What it does

Dyne is a web application that helps people decide where to eat through a simple swipe-based experience. Users swipe through nearby restaurants, and Dyne uses those interactions to quickly surface options that align with their preferences. Instead of endless scrolling or debating, users decide in seconds.

How we built it

Dyne was built as a responsive web application designed for speed and simplicity. We use location-based restaurant data to surface nearby options and a swipe-based interface to collect clear preference signals. These signals are stored per user and fed into a lightweight machine-learning model that continuously improves restaurant ranking over time.

Challenges we ran into

• Deciding which features actually mattered
• Avoiding feature bloat that slowed down the core decision flow
• Staying focused on a single action that helps users decide quickly

Accomplishments that we're proud of

• Shipped a fully functional end-to-end web application during the hackathon
• Implemented a lightweight ML model that learns from swipe behavior
• Improved restaurant ranking over time with continued user interaction
• Created a clean, intuitive UI that requires no onboarding

What we learned

Users don’t want to evaluate options—they want a decision. When given too many details or controls, people slow down instead of moving forward. Simple inputs like swipes produce better outcomes than complex filters. We also learned that small, repeated signals are enough to drive useful recommendations when the product stays focused. Cutting features and committing to a single clear loop helped us move faster and build something people immediately understand.

What's next for Dyne

• Add group decision-making so friends can swipe together
• Help groups quickly agree on a restaurant
• Continue improving the learning model for better recommendations
• Reduce the number of swipes needed over time
• Introduce social features like chat and photo sharing
• Let users share food experiences and engage with others

Built With

• azure
• figma
• javascript
• next.js
• python
• scikit-learn
• tailwind