# Skill Card

## Description

Autonomously optimizes end-to-end LLM inference throughput on AMD Instinct GPUs with Hyperloom and reports a validated gain.

## Owner

AMD AGI / Hyperloom

## License

MIT