Max Krasnyansky
Author of llama.cpp's Hexagon NPU backend; member of its Hexagon and OpenCL code-owner teams
- Member of the Hexagon and OpenCL backend code-owner teams, llama.cpp project2
Max Krasnyansky opened the pull request that added ggml-hexagon, an experimental llama.cpp backend for the Hexagon NPU, which was merged in October 2025. llama.cpp's CODEOWNERS file lists Krasnyansky in the teams that own the Hexagon and OpenCL backends; the Hexagon team also owns the project's Snapdragon documentation and scripts. Krasnyansky's 2026 commits extend the Hexagon backend with multi-NPU and multi-device support, reworked DMA, buffer and cache handling, matrix-multiplication and FlashAttention updates, and a backend sampler.123
Work
llama.cpp Hexagon backend (ggml-hexagon)Open-source code
Opened pull request #16547, "Add experimental ggml-hexagon backend for the Hexagon NPU," merged in October 2025. CODEOWNERS lists Krasnyansky in the ggml-org/ggml-hexagon and ggml-org/ggml-opencl teams. Commits authored from July to October 2026 add multi-NPU and multi-device support, DMA and L2-cache rework, MUL_MAT and FlashAttention updates, and backend sampler support. The repository is MIT-licensed.1234
This profile covers public professional work only, from the sources below. It is not an endorsement or a ranking. To correct or remove it, use the correction form.