AI People Wiki

Builders & open source

Adrien Gallouët

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

  • Hugging Face is the company listed on Gallouët's GitHub profile, Hugging Face12
  • Member of the llama-server and llama-common code-owner teams, llama.cpp project3

Adrien Gallouët contributes to llama.cpp, the C/C++ runtime for running language models on local hardware; Gallouët's GitHub profile lists Hugging Face as company and shows membership in the huggingface and ggml-org organizations. llama.cpp's CODEOWNERS file includes Gallouët in the teams that own llama-server, the project's HTTP server, and its shared common code. Gallouët's recent commits add model input and output modalities to the server's model listing, rework file-path and cache-directory handling across platforms, clean up compiler warnings in the CUDA and Windows builds, and update the vendored HTTP library.1234

At a glance

Current roles
Hugging Face is the company listed on Gallouët's GitHub profile +1 more
Sector
Company
Work listed
3
Sources cited
12
Profile reviewed

Open-source work

Adrien Gallouët's pinned GitHub projects center on networking and systems software, mostly in C. They include Glorytun, a multipath UDP tunnel, and MUD, the multipath UDP library behind it, along with a small secret store, an X11 cloud desktop tool, a minimal PID 1 for initramfs, and an HTTPS API for the Slurm workload manager.156

Glorytun's README describes encrypted, authenticated, and timestamped messages, forward secrecy with ECDH over Curve25519, multipath aggregation with backup paths, and traffic shaping. MUD's README notes that it runs on Linux and has been ported to OpenBSD and macOS.56

llama.cpp

A November 2024 llama.cpp commit by Gallouët replaced AArch64 NEON assembly with intrinsics in a quantized matrix-vector routine. Gallouët's recent commits span the CPU and CUDA backends, the shared common library, the HTTP server, and tools such as llama-bench.74

In June 2026 Gallouët added a llama download subcommand, and in October 2026 made the server report each model's input and output modalities in GET /models. CODEOWNERS now places Gallouët on the teams responsible for the server and the common code.893

Gallouët also started installama.sh as a quick way to install llama.cpp. The project became an official llama.cpp package in the ggml-org organization as llama-install.sh, which provides install scripts for Linux, macOS, FreeBSD, and Windows.1011

Work and contributions

  • llama.cpp Open-source code

    CODEOWNERS lists Gallouët in the ggml-org/llama-server and ggml-org/llama-common teams. Commits authored in September and October 2026 report model modalities in the server's GET /models endpoint, move common code to fs::path for config and cache directories, clean up CUDA and Windows warnings, and update the vendored cpp-httplib. The repository is MIT-licensed.3412

  • llama-install.sh (formerly installama.sh) Open-source code

    Started installama.sh, a script for installing llama.cpp, which has since migrated to the ggml-org organization as llama-install.sh. The MIT-licensed scripts download a prebuilt llama binary and detect CPU and GPU acceleration such as CUDA, ROCm, Vulkan, and Metal.1011

  • Glorytun and MUD Open-source code

    Wrote Glorytun, a small multipath UDP tunnel, and MUD, the secure multipath network protocol over UDP that provides Glorytun's core features, including AEGIS-256 or ChaCha20-Poly1305 encryption and link aggregation with failover.56

Timeline

  1. Authored a llama.cpp commit replacing AArch64 NEON assembly with intrinsics.7
  2. Added the llama download subcommand to llama.cpp.8
  3. Made llama.cpp's server report model input and output modalities in GET /models.9

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

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

  1. angt (Adrien Gallouët) on GitHubGitHub · Official page · accessed Oct 21, 2026
  2. angt on Hugging Face (organizations include huggingface and ggml-org)Hugging Face · Official page · accessed Oct 21, 2026
  3. ggml-org/llama.cpp CODEOWNERSggml-org (GitHub repository) · Repository · accessed Oct 21, 2026
  4. ggml-org/llama.cpp commits authored by angtggml-org (GitHub repository) · Repository · accessed Oct 21, 2026
  5. angt/glorytun READMEangt (GitHub repository) · Repository · accessed Oct 21, 2026
  6. angt/mud READMEangt (GitHub repository) · Repository · accessed Oct 21, 2026
  7. ggml-cpu: replace AArch64 NEON assembly with intrinsics in ggml_gemv_q4_0_4x4_q8_0() (#10567)ggml-org (GitHub repository) · Repository · published Nov 30, 2024 · accessed Oct 21, 2026
  8. app : add the llama download subcommand (#24982)ggml-org (GitHub repository) · Repository · published Jun 25, 2026 · accessed Oct 21, 2026
  9. common, server : report model input/output modalities in GET /models (#29987)ggml-org (GitHub repository) · Repository · published Oct 5, 2026 · accessed Oct 21, 2026
  10. angt/installama.sh READMEangt (GitHub repository) · Repository · accessed Oct 21, 2026
  11. ggml-org/llama-install.sh READMEggml-org (GitHub repository) · Repository · accessed Oct 21, 2026
  12. ggml-org/llama.cpp LICENSE (MIT)ggml-org (GitHub repository) · License · accessed Oct 21, 2026