Developed a SwiftUI iOS application enabling users to track caffeine and calorie intake across 15+ customizable beverages. Leveraged MVVM architecture with ObservableObject for reactive state management, persistent data storage, and dynamic nutritional calculations based on serving size, milk, and syrup options.

Built an end-to-end retrieval-augmented generation system for private datasets, delivering context-aware answers with low latency. Implemented text chunking, embeddings, vector search, and evaluation tooling for observability.