Hugging Face Releases ml-intern: An Open-Source AI Agent that Automates the LLM Post-Training Workflow

Hugging Face released ml-intern, an open-source agent built on smolagents to automate LLM post-training workflows.

It can search papers and datasets, launch training jobs, track experiments, and retry after failures with less manual input.

In demos, it raised Qwen3-1.7B from about 10% to 32% on GPQA in roughly 10 hours using methods like synthetic data and GRPO.

The launch highlights how agentic systems may speed up post-training and improve benchmark results for small models.