←Leaderboard/Developer Tooling/inference-optimisation-skill
Last commit on November 16, 2026·Created on November 16, 2026

sybil-solutions/inference-optimisation-skill

“Optimize the serving speed of local LLMs.”
Combined rank
#574
across all skills
In Developer Tooling
#130
category rank
Stars
564
7d change unavailable
Forks
57
Watchers
2
Traction scoreGitHub stars can be faked, so popularity alone can be misleading. Traction Score looks for broader signs of recent attention, adoption, and active maintenance.
TL;DR

This repository provides a methodology for hill-climbing the performance of model servers across various hardware and quantization formats.

WHO IT'S FOR
MLOps / Infrastructure engineers
optimise local LLM serving speed and throughput
AI Performance engineers
benchmark speculative decoding and prefill speeds
Hardware / Systems architects
validate LLM performance across diverse hardware targets
Quantization researchers
evaluate performance trade-offs for different quantization formats
Repository contents

1 skill files

Compatible AgentsThe repository documents support for these agents. The skills may also work with other agents that can load SKILL.md files, but they may need some setup or small changes.

Use this skill

Follow the documented setup, then try a first task.

I’m using

Claude Code user-level installation

Installs: User-level installation of inference-optimisation. · Claude Code

Before you start
  • Git must be available to clone the repository.
  1. Run this complete block in a terminal to clone the repository and copy the skill into Claude Code's user skills directory.

    git clone https://github.com/sybil-solutions/inference-optimisation-skill
    mkdir -p ~/.claude/skills
    cp -r inference-optimisation-skill/skills/inference-optimisation ~/.claude/skills/
README.md ↗ · Checked Nov 17, 2026

Give it something to do.

Suggested first task

Estimate your model's theoretical speed ceiling

Uses inference-optimisation

  • The model and its quantization.
  • Access to the target machine or its hardware specifications.
Use the inference-optimisation skill to estimate the theoretical maximum tokens per second for my model on this machine. Inventory the machine and explain the assumptions behind the estimate.
README.md ↗ · Checked Nov 17, 2026skills/inference-optimisation/SKILL.md ↗ · Checked Nov 17, 2026