This repository contains the algorithm designs.

Designed to:

• Show user profiles strategically instead of randomly
• Balance between familiar (same university or degree) and diverse profiles
• Learn from user like/dislike patterns to improve recommendations
• Keep users engaged while maximising meaningful connections

Designed to:

• Form balanced dinner groups of six people with compatible dietary, budget, and location constraints
• Create groups with complementary interests to encourage natural conversations
• Ensure fairness so every user gets equal opportunities for great experiences

Both algorithms are designed to be:

• Scalable to 10,000+ users per city
• Fast, executing within 100ms
• Adaptable to learn and improve over time
• Fair and privacy-conscious

Detailed explanations, logic, trade-offs, and implementation plans are included in their respective files.