WecoAI/weco-cli

Weco helps engineers and researchers automatically improve their code, specifically for tasks like optimizing GPU operations, refining machine learning models, or enhancing large language model prompts. You provide your existing code and an evaluation script that outputs a performance metric (e.g., latency, accuracy, win rate). Weco then iteratively modifies your code to maximize or minimize that metric, delivering a more performant version of your original code. This is for machine learning engineers, GPU developers, and prompt engineers looking to boost code efficiency or model quality.