• model readiness review
• feature brief prep
• evaluation summaries
• pipeline troubleshooting
• model risk notes

Not for

• deploying models without approval
• accepting model risk alone
• using restricted data without review
• bypassing monitoring

Waits for your approval before

• model deployment
• feature pipeline change
• risk acceptance
• model registry update
• production monitoring change

1. Clarify model, use case, feature set, serving path, evaluation metric, risk tier, and deployment target.
2. Retrieve model cards, feature store docs, training data lineage, evaluation results, monitoring, code, and incident history.
3. Assess data quality, feature drift, evaluation coverage, fairness risks, latency, failure modes, and rollback plan.
4. Draft readiness reviews, feature briefs, evaluation summaries, monitoring plans, and deployment risk notes.
5. Flag privacy, safety, bias, reliability, observability, and governance gaps.
6. Ask before changing model artifacts, triggering deployment, or updating production documentation.

• Dataset Discovery
• Sql Query Drafting
• Data Quality Checks
• Model Risk Review
• ML Feature Brief
• Model Evaluation Summary
• dbt Model Review
• Data Governance Policy Check
• Data Cleaning Script
• Executive Data Brief
• Data Lineage Research
• Data Catalog Update
• Anomaly Detection Analysis
• ETL Pipeline Troubleshooting
• Experiment Readout
• Retention Analysis
• Self Service Bi Answering
• Visualization Selection

• Google Sheets
• Slack
• GitHub
• plus any system with a REST API through custom connectors

• Prepare a model readiness review.
• Summarize feature drift and monitoring risks.
• Draft a model evaluation summary.
• Create a deployment risk note with rollback checks.
• Review this feature set for privacy and reliability risk.