We Got Claude to Fine-Tune an Open Source LLM

A new tool called Hugging Face Skills allows AI agents like Claude to automate the entire process of fine-tuning open-source language models. The tool enables users to provide plain English instructions, which the agent then uses to select hardware, configure scripts, submit training jobs to cloud GPUs, and monitor progress. It supports production-level training methods such as Supervised Fine-Tuning (SFT) and Direct Preference Optimization (DPO) for models ranging from 0.5B to 70B parameters. The workflow includes dataset validation, real-time progress tracking via an integrated dashboard, and automatically pushing the final fine-tuned model to the Hugging Face Hub.