Data Quality Agent
Continuously monitors data quality and quarantines bad data before it spreads.
Profiles every feed, learns a per-feed baseline of 'normal', and detects schema drift, unit changes and null-rate spikes against it. It quarantines the bad batch before it reaches a risk model or regulatory report and notifies the feed owner with the diff and suspected root cause.
Authority
Act within policy
Team role
Monitors and escalates
Handoffs
Named collaborators
The role
What it owns and where its authority ends
Desk
Data Governance & Quality
Desk workflow
Ingest, then profile, then quality checks, then lineage capture, then catalog and certify. Bad batches are quarantined before they reach downstream models, reports and agents.
Collaboration
Coordinates parallel monitors or workers
Decision boundary
Acts only within an explicit policy, permission and escalation boundary.
Systems and capabilities involved
Data pipeline / warehouse API
read batches, quarantine
Profiling + anomaly sandbox
Data-contract registry
Owner notification channel
Handoffs
What this role gives and receives
Capabilities offered
The handoffs name the next owner or specialist and the work that moves between them.
Handoff to
External handoff
Data owners in Markets and Risk for upstream fixes
Context
What the role needs to do the work
- Current work
- The current batch profile vs. the learned baseline.
- Prior interactions
- Prior quality incidents per feed and their root causes.
- Policies and reference
- Data contracts, expected schemas, business validation rules.
- Working method
- Not specified for this role.
Illustrative workflow
How the work moves
Starting point
Nightly market-data feed loads; one instrument's prices arrive 100x expected.
- 01
Profile the batch; detect a 100x scale anomaly on a single symbol vs. baseline.
- 02
Confirm it's a unit error (cents vs. dollars), not a real move, via cross-source check.
- 03
Quarantine the affected rows before they hit the valuation model.
- 04
Notify the feed owner with the diff and suspected root cause.
Result
Bad batch quarantined pre-ingestion; the risk model runs on clean data; the feed owner receives a root-caused alert with the diff.
Checks and boundaries
What must be tested or reviewed
- 01Quarantine actions are reversible and logged; false-quarantine rate tracked.
- 02Drift detection on the baseline itself: alert if 'normal' shifts too fast.
- 03Critical regulatory feeds (BCBS 239) require an oversight-agent gate to release after quarantine.
Human authority
Acts only within an explicit policy, permission and escalation boundary.
Keep exploring