MLE Agents: Autonomous ML Engineering in Large-scale Recommendation Systems
Organizations: Meta Platforms
Abstract
The typical workflow for improving large-scale ranking and recommendation models is slow and tedious: an engineer forms a hypothesis, implements it as a code change, trains a model, and evaluates it both offline and online, often relying on their own intuition and experience along the way. Autonomous LLM agents promise to speed this up by running the loop on their own, but it is still unclear whether they can do this in industry-scale systems rather than on benchmarks or small proxy tasks. We build human-on-the-loop machine learning engineer (MLE) agents that propose, implement, train, and evaluate model changes on large-scale search rankers. Engineers retain control over objective design, search direction, and final-model review. We identify three obstacles that make this industry setting challenging: feedback takes hours per experiment, offline metrics do not always predict online impact, and improvements saturate after the early wins. We address these with parallel training runs, an objective calibrated to match online impact, and careful context engineering. In one campaign, the system ran 1,097 completed candidate experiments and accepted 325, a 29.6% acceptance rate compared with 15.6% in the earlier campaign under a different objective and search constraint. Its best model increases search volume by 0.23% in an A/B test of large scale. Several improvements were architectural changes the team had never tried, drawn from the broader industry and literature, suggesting that MLE agents can explore regions of the design space that human intuition overlooks. We find that the design of the harness around the model matters: intuition suggests that a stronger underlying model is enough, but in our experience, the design of the loop around it is just as critical. More broadly, our work shows that autonomous agents can do real machine-learning engineering in industry-scale systems.
Figures & tables
| Change (agent-generated) | weighted NE |
|---|---|
| content_downstream neg_label_smoothing 0.020.005 | -0.0041 |
| Remove content_downstream neg_label_smoothing; content_downstream loss_weight 2.5 | -0.0038 |
| Learned sigmoid gate on attention output (pre-FFN) in Self-Attention transformer blocks | -0.0024 |
| Deeper attentive-distillation MLP [1024,512][1024,768,512], zero-init residual | -0.0021 |
| Gated residual bypass from attentive-distillation output into task towers | -0.0021 |
| Self-Attention transformer 43 blocks with FFN expansion 14 | -0.0020 |
| Online metric | Impact |
|---|---|
| Search session volume | |
| Downstream engagement (participation, retention) | Directionally positive |
Automated Machine Learning
Automated machine learning (AutoML) builds and tunes data-to-model pipelines with minimal manual effort, balancing predictive quality, compute, and deployment constraints.
See all 105 papers