MLExpert: Towards a Generalist Agentic Model for Autonomous Machine Learning Engineering
35B MLExpert raises MLE-Bench Lite medal rate from 37.9% to 72.7%.
The paper introduces a post-training method for autonomous, long-horizon machine-learning engineering and uses it to train MLExpert, a 35B Mixture-of-Experts agentic model. Hierarchical agentic supervised fine-tuning and Windowed Interaction RL learn from intermediate experiment outcomes, with asynchronous collection enabling updates before a trajectory ends. On MLE-Bench Lite, MLExpert raises the base model’s average medal rate from 37.9% to 72.7%, comparable to larger models such as Kimi-K3 and GLM-5.3. It also improves NatureBench Lite, agent-scaffold optimization, and algorithm optimization.
- MLExpert is a 35B Mixture-of-Experts agentic model for long-horizon MLE.
- Hierarchical SFT and Windowed Interaction RL use outcome-based rewards.
- MLE-Bench Lite medal rate rose from 37.9% to 72.7%.
- Results approach Kimi-K3 and GLM-5.3 and transfer to other optimization tasks.
Full article247 words · extracted from arxiv.org · click to collapse
Large language models (LLMs) are expanding from assisting scientific research and engineering to autonomously developing and improving AI systems, advancing AI for AI (AI4AI) as a pathway toward recursive self-improvement (RSI). Workflow-driven machine learning engineering (MLE) systems provide a practical path toward this pursuit through complex orchestration that manages long-horizon experimentation, while recent MLE systems adopt model-directed interaction, allowing models to autonomously organize experiments and adapt optimization strategies based on execution feedback. However, feedback from each experiment reflects several preceding decisions, and a long MLE run contains many such costly experiments. Training must therefore select useful supervision from long interaction trajectories and learn from intermediate outcomes. We propose a post-training method for autonomous long-horizon MLE optimization and apply it to train MLExpert, a 35B Mixture-of-Experts agentic model. We design the Kaggle-MLE Factory to construct executable tasks and an interaction harness for model-directed experimentation. Hierarchical agentic SFT transfers general tool-use and coding capabilities to MLE through outcome-guided supervision. Windowed Interaction RL (WIRL) treats the multi-turn decisions between successive experiment results as an execution window and assigns it a shared outcome-based reward. Asynchronous collection and policy correction support policy updates before the trajectory ends. On MLE-Bench Lite, MLExpert improves the average medal rate of the base model from 37.9% to 72.7%, achieving performance comparable to larger frontier models such as Kimi-K3 and GLM-5.3. Further experiments show that MLExpert improves performance on NatureBench Lite, agent scaffold optimization, and algorithm optimization, suggesting that the learned optimization capability generalizes to other tasks.
Text extracted automatically; images, tables and formatting may be missing. Original: https://arxiv.org/abs/2610.22344