Platform Guides

ML Operations

Register models, run experiments and evaluations, launch training jobs, and monitor models in production.

Overview

ML operations cover the full model lifecycle. Register a model, run experiments to compare configurations, score them with evaluations, kick off training jobs, and watch behavior over time with monitoring. Curated label data leaves the platform via ML data exports.

Models & experiments

GET/api/v1/ml-modelsList registered ML models.
POST/api/v1/ml-modelsRegister an ML model.
GET/api/v1/ml-experimentsList experiments.
POST/api/v1/ml-experimentsCreate an experiment.

Evaluations & training

GET/api/v1/evaluationsList model evaluations.
POST/api/v1/evaluationsRun an evaluation.
POST/api/v1/training-jobsLaunch a training job.
GET/api/v1/training-jobs/{id}Check training job status.

Monitoring & exports

GET/api/v1/monitoringRead production monitoring signals.
GET/api/v1/performance-metricsList model performance metrics.
POST/api/v1/ml/data-exportsExport curated labeled data.

For agent-specific evaluation and observability, see the AEOS overview. Full schema in the Label Backend API reference.