My path in software engineering is rooted in a relentless drive for bare-metal hardware control, performance optimization, and architectural cleanliness. Rather than following the prevailing industry trend of packaging sluggish web wrappers into memory-hungry runtimes, I specialize in lean, native code built with Rust, C++, and modern Python — communicating directly with GPU hardware via CUDA, Tensor Cores, and Vulkan NCNN. Every CPU clock cycle, optimal VRAM memory access pattern, and sub-millisecond UI response time matters.
I firmly believe that the future of artificial intelligence belongs to the Local-First paradigm. In an era dominated by subscription fatigue, fragile cloud dependencies, and invasive data telemetry, I engineer tools that deliver 100% mathematical privacy, zero network latency, and permanent lifetime software ownership. Your high-resolution media, personal photos, and sensitive documents should never become data points on a remote server — all neural compute runs completely on your own local workstation.
As both an engineer and a physical maker, I bridge the gap between digital software pipelines and real-world mechanical prototyping. Alongside desktop development, I operate a dedicated hardware lab: designing parametric assemblies in Autodesk Fusion 360, building and calibrating custom high-speed CoreXY 3D printers, tuning Klipper/Moonraker firmware, and manufacturing functional parts from engineering composites. This direct fusion of low-level software and mechanical physics yields tools that are exceptionally robust, practical, and engineered for real-world demands.