I build projects around constrained hardware, low-level performance, and practical ML deployment, mostly in C/C++, Python, and Kotlin.
Main projects: llmedge, LightDiffusion-Next, EasyReader.
- native and mobile AI runtimes
- efficient inference for LLMs and diffusion models
- performance-oriented systems design
- embedded / real-time oriented projects
Languages: C, C++, Python, Kotlin, Rust
Focus: PyTorch, mobile/edge inference, Vulkan/OpenCL, embedded systems, signal processing