Development of a User-Centric Digital Menu with Adaptive Personalization System

This thesis presents the conception and development of SpaceM, a comprehensive digital menu solution operating within the WaitHero ecosystem. Designed to address the limitations of earlier implementations, SpaceM introduces persistent user profiling, multi-method authentication, and user-centered customization features, enabling personalized dining experiences for a diverse and growing user base. By capturing individual preferences such as dietary restrictions, allergens, and favorite items, the platform tailors both the interface and the recommendations presented to each customer. A core technical challenge lies in reconciling disparate product data generated by numerous restaurants. To handle this, SpaceM employs a layered relational schema connecting global “master” products with restaurant-specific entries. Additionally, a human-validated database underpins an automated NLP-driven clustering workflow capable of handling hundreds of thousands of product records. In parallel, a recommendation engine—implemented in Rust—performs nightly batch processing to aggregate and interpret user interactions, producing real-time ranking scores that power the “Chosen For You” feature. This hybrid design, combining offline analysis with real-time filtering, accommodates fine-grained user preferences and allergen constraints while ensuring swift menu rendering. Looking ahead, the thesis outlines strategies for boosting system scalability: transitioning static assets to AWS S3, adopting distributed caching (Memcached clusters), and leveraging Cassandra for seamless expansion of user tracking data. The result is a robust, easily extensible platform capable of meeting the evolving needs of modern food-service operations and paving the way for further innovations in personalized hospitality technologies.