Native Mobile Developer / Local AI Enthusiast
I build intelligent, deeply integrated native applications that prioritize on-device AI, low-latency performance, and strict privacy. I work across modern Android architecture and the Apple ecosystem (macOS/Swift). I focus on minimalist precision—while ensuring heavy models run fast and cool on fanless hardware.
Jarvis (macOS Native) (https://github.com/Sam-06060/jarvis-assistant) — A privacy-first, voice and text controlled personal assistant. Runs entirely on online llms and can load llms locally on-demand and when offline, leveraging the Apple Neural Engine for zero-cloud dependency and optimal thermals. Features a custom hybrid STT engine (Apple Speech + Whisper) for instant transcription, "Mimic" macro recording, "DeadDrop" file transfers, and secure Face ID authentication.
Tree Species Detection (Android) — Mobile client-server application for local flora identification. Integrates a modern Android UI with a Railway-hosted backend for heavy image analysis, featuring seamless API integration and robust local CSV data export functionality.
Local AI & Edge Compute — Optimizing LLM inference (Ollama) for constrained and fanless architectures. Building ultra-low latency hybrid voice pipelines. Managing thermal overhead by intentionally avoiding heavy tracking libraries and routing compute efficiently.
Production Mobile Architecture — Android MVVM, local state persistence via RoomDB, secure biometric integration. macOS native development in Swift, interacting with low-level system APIs for deep OS automation.
High-Fidelity UI/UX — Translating complex logic into minimalist interfaces. Deep practical knowledge of responsive visual feedback loops, and creating seamless, fluid animations that mimic first-party OS experiences.
- Refining my personal macOS Desktop assistant Jarvis with new features and enhancements.
- Deepening architectural patterns (MVVM, RoomDB) for production Android environments.