AI People Wiki

Builders & open source

Sigbjørn Skjæret

Hugging Face is the company listed on Skjæret's GitHub profile, Hugging Face

  • Hugging Face is the company listed on Skjæret's GitHub profile, Hugging Face1
  • Code owner for conversion, gguf-py and core model code; member of the CI team, llama.cpp project2

Sigbjørn Skjæret, who publishes on GitHub as CISC, contributes to llama.cpp, the C/C++ runtime for running language models on local hardware; Skjæret's GitHub profile lists Hugging Face as company and shows membership in the huggingface and ggml-org organizations. In llama.cpp's CODEOWNERS file, Skjæret owns the model conversion scripts, the gguf-py Python package, the core model, architecture, graph and vocabulary code, and the Jinja template code in common, and is a member of the CI team. Skjæret's recent commits cover the project's CI and release workflows, Jinja template support, GGUF tokenizer handling, and model fixes.123

At a glance

Current roles
Hugging Face is the company listed on Skjæret's GitHub profile +1 more
Sector
Company
Work listed
1
Sources cited
4
Profile reviewed

Open-source work

Skjæret, who works on GitHub as CISC, is a code owner in llama.cpp, the MIT-licensed C/C++ project for running language models on local hardware. The project's CODEOWNERS file assigns Skjæret ownership of the Python model conversion scripts and the gguf-py package, of core model code including the architecture, graph, vocabulary, adapter and per-model source files, of the Jinja template code in common, and of the Python requirements files. Skjæret is also listed in the project's CI team.241

Skjæret's commits from September and October 2026 show much of this work. They include refactoring the continuous integration and release workflows, adding a Windows Arm64 Vulkan release build, extending the built-in Jinja template engine with features such as dict support and new test statements, refining how gguf-py handles tokenizer special tokens and normalization, adding fill-in-the-middle tokens for PLaMo models, and a CUDA fusion for grouped experts top-k selection.3

Work and contributions

  • llama.cpp Open-source code

    CODEOWNERS names Skjæret for /conversion/, convert_*.py, gguf-py, common/jinja, the requirements files, and src/llama-arch, llama-model, llama-graph, llama-vocab, llama-adapter and src/models, and lists Skjæret in the ggml-org/ci team. Commits authored in September and October 2026 cover CI and release workflows, Jinja features, gguf-py tokenizer handling, and model fixes. The repository is MIT-licensed.234

Timeline

  1. Refactored llama.cpp's self-hosted CI builds into backend-specific workflows.3
  2. Added support for the dict builtin to llama.cpp's Jinja template engine.3
  3. Improved llama.cpp's release flow and added a Windows Arm64 Vulkan release.3
  4. Added detection of NFC normalization to the gguf-py package.3
  5. Added a grouped experts top-k fusion to llama.cpp's CUDA backend.3

Where they work

Suggest a correction or removal

Profiles cover public professional work only, and each statement cites a source. If something is out of date, missing or wrong, tell us and link a source. People can also ask for their profile to be corrected or removed.

References

4 sources · profile reviewed Oct 21, 2026 · data updated Oct 21, 2026

  1. CISC (Sigbjørn Skjæret) on GitHubGitHub · Official page · accessed Oct 21, 2026
  2. ggml-org/llama.cpp CODEOWNERSggml-org (GitHub repository) · Repository · accessed Oct 21, 2026
  3. ggml-org/llama.cpp commits authored by CISCggml-org (GitHub repository) · Repository · accessed Oct 21, 2026
  4. ggml-org/llama.cpp LICENSE (MIT)ggml-org (GitHub repository) · License · accessed Oct 21, 2026