Learning to Ideate for Machine Learning Engineering Agents

TL;DR

This paper introduces MLE-Ideator, a dual-agent framework that separates ideation from implementation in machine learning engineering, significantly improving effectiveness through training-free and RL-trained approaches.

Contribution

The paper proposes a novel dual-agent system for MLE that enhances algorithm optimization by separating ideation from implementation, with effective training strategies.

Findings

Outperforms implementation-only baselines on MLE-Bench

RL-trained Ideator achieves 11.5% improvement with limited data

RL training surpasses existing models like Claude Sonnet 3.5

Abstract

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.