MLflow experiment name

MLflow run name

MLflow log_params

MLflow log_metrics

MLflow log_model with registered name

Evaluation summary task description

Evaluation summary dataset info

Evaluation summary baseline vs candidate

Latency in evaluation summary

Slice evaluation present

Failure mode identified

Deployment recommendation

Stratified split used

No fabricated numbers in prose

FastAPI used

Pydantic request schema

Pydantic response schema

response_model used

POST /predict endpoint

Health check endpoint

MLflow model registry URI

mlflow.sklearn.load_model used

numpy array reshape

Score extraction

Checklist: held-out metrics

Checklist: slice evaluation

Checklist: artifact versioning

Checklist: smoke test

Checklist: latency benchmark

Checklist: rollback artifact

Checklist: monitoring dashboards

Checklist: rollback runbook

Checklist: privacy controls

Drift metric: PSI threshold

Retraining trigger: performance drop

Retraining trigger: data volume

Rollback trigger: error rate

Rollback trigger: latency duration