Existing machine learning engineering (MLE) agents struggle to iteratively optimize their implemented algorithms for effectiveness. To address this, we introduce MLE-Ideator, a dual-agent framework that separates ideation from implementation. In our system, an implementation agent can request strategic help from a dedicated Ideator. We show this approach is effective in two ways. First, in a training-free setup, our framework significantly outperforms implementation-only agent baselines on MLE-Bench. Second, we demonstrate that the Ideator can be trained with reinforcement learning (RL) to generate more effective ideas. With only 1K training samples from 10 MLE tasks, our RL-trained Qwen3-8B Ideator achieves an 11.5% relative improvement compared to its untrained counterpart and surpasses Claude Sonnet 3.5. These results highlights a promising path toward training strategic AI systems for scientific discovery.
@article{arxiv.2601.17596,
title = {Learning to Ideate for Machine Learning Engineering Agents},
author = {Yunxiang Zhang and Kang Zhou and Zhichao Xu and Kiran Ramnath and Yun Zhou and Sangmin Woo and Haibo Ding and Lin Lee Cheong},
journal= {arXiv preprint arXiv:2601.17596},
year = {2026}
}