Swift, Rust, Tauri, Docker, AWS

Built a desktop-native multi-modal LLM platform with prompt branching and multi-window PIP workflows.
Integrated OpenAI, Claude, and Gemini APIs to run prompts across models for comparison and reliability analysis.
Implemented orchestration layers to manage prompt routing, response aggregation, and concurrent model execution.
Developed native macOS client in Swift and a Windows client using Tauri + Rust optimized for resource efficiency.
Containerized backend with Docker, deployed via AWS App Runner, and orchestrated components using Kubernetes.