Data governance specialist

What does a data governance specialist do?

A data governance specialist sets up and runs the rules for how an organisation collects, stores, shares and uses data.

You make sure data is accurate, easy to find and handled in line with company policy and the law.

The role usually sits in a data office, a data and analytics team, or a risk and compliance function. It is common in banks, insurers, government agencies, healthcare and large retailers. These hold a lot of personal data covered by the Personal Data Protection Act (PDPA).

You work with data engineers, analysts, data architects, legal and compliance staff, information security teams and the business units that own the data.

What the work involves

• Write and update data policies, standards and definitions, and keep them in a shared wiki such as Confluence or Notion.
• Run data quality checks in SQL or Great Expectations, then follow up with the owners of failed checks.
• Keep the data catalogue and business glossary current so staff can find trusted datasets.
• Review requests to share or use personal data, and check them against PDPA obligations and internal rules.
• Meet data owners and stewards to agree who is responsible for each dataset and to clear open issues.
• Report data quality and policy compliance results to a data governance committee.

Pay in Singapore

Back to top

Skills a data governance specialist needs

Back to top

Data governance specialist job description

We are hiring a data governance specialist to help us manage data as a shared asset.

You will maintain data policies and standards, keep the data catalogue up to date, and run data quality checks. You will also support compliance with the PDPA. You will work with business owners, engineers and legal teams across the organisation.

Responsibilities

• Draft, maintain and publish data policies, standards and procedures.
• Define and record business terms, data owners and data stewards for key datasets.
• Maintain the data catalogue and record where data comes from and where it goes.
• Set up data quality rules and tests, and track issues until they are fixed.
• Classify data by sensitivity and check that access and retention follow the classification.
• Review projects and vendor requests that involve personal data, and flag PDPA risks.
• Support data protection impact assessments and the response to data incidents.
• Assess governance risks when teams build or buy AI and machine learning tools.
• Run training and briefings so staff know their data responsibilities.
• Report progress and open risks to the data governance committee and senior management.

Requirements

• Degree or diploma in information systems, computer science, data analytics, business or a related field. Some employers accept law or risk backgrounds.
• Typically 2 to 5 years in data analysis, data management, risk, audit or compliance.
• Working SQL for profiling data, tracing records and checking quality.
• Good knowledge of the PDPA and how it applies to day-to-day data handling.
• Experience with documentation tools such as Confluence or Notion, and ideally a data catalogue or data quality tool.
• Clear writing and the ability to explain data rules to non-technical colleagues.

Back to top

Data governance specialist career path

Data governance careers in Singapore often start in data analysis, risk or compliance and move towards policy ownership and leadership.

The titles below are common, but they vary by employer.

1. Data governance analyst

   Entry

   You support the team by documenting data definitions, running quality checks and updating the catalogue under guidance.

2. Data governance specialist

   Mid

   You own policies and standards for a set of data domains, work directly with data owners and handle PDPA reviews.

3. Senior data governance specialist

   Senior

   You lead governance for large programmes, advise on AI and data risk, and mentor junior staff.

4. Data governance manager

   Lead

   You run the governance team, chair working groups and set priorities across business units.

5. Head of data governance or chief data officer

   Head

   You set the data strategy and governance model for the organisation and report to senior management or the board.

The titles are ones commonly used in Singapore; employers may name the levels differently.

Related jobs and their pay

Back to top

Data governance specialist interview questions

We wrote these questions from the skills this job needs. They are not an official question bank.

1. What is the difference between a data owner, a data steward and a data custodian?

   What a good answer covers

   • Owner: a senior business person accountable for a dataset and who may use it.
   • Steward: handles day-to-day definitions, quality and issue resolution.
   • Custodian: the technical team that stores and secures the data.
   • Gives an example of how the three roles work together on one dataset.

2. Write a SQL query to find customers with duplicate email addresses and keep only the latest record for each.

   What a good answer covers

   • Uses GROUP BY with HAVING COUNT greater than 1 to find duplicates.
   • Uses ROW_NUMBER over a partition by email, ordered by update time, to pick the latest record.
   • Normalises email case and whitespace before comparing.
   • Says a business rule is needed to decide which record is correct, since the latest date may not be the right one.

3. Which data quality dimensions would you measure for a customer table, and how would you test them in Great Expectations?

   What a good answer covers

   • Names dimensions such as completeness, uniqueness, validity, timeliness and consistency.
   • Maps each to an expectation, for example not-null, unique, value in set or regex match.
   • Sets thresholds with the data owner instead of choosing them alone.
   • Describes how results are stored, alerted on and reported.

4. How would you build a business glossary and data catalogue for a company that has neither?

   What a good answer covers

   • Starts with the most critical datasets and reports, then widens over time.
   • Agrees definitions with business owners and records them together in the catalogue.
   • Captures owner, steward, source system and sensitivity for each entry.
   • Sets a review cycle so entries do not go out of date.

5. How would you classify data by sensitivity, and what controls follow from each level?

   What a good answer covers

   • Defines a small number of levels, for example public, internal, confidential and restricted.
   • Places personal data and national identification numbers in the higher levels.
   • Links each level to access, masking, encryption and retention rules.
   • Explains how classification is applied to new datasets.

6. How would you organise data policies in Confluence or Notion so staff actually use them?

   What a good answer covers

   • Uses a clear structure with a single page per policy and a named owner.
   • Adds plain-language summaries and examples next to formal wording.
   • Shows version history, review dates and a way to ask questions.
   • Tracks which pages are read or cited and removes outdated ones.

7. What is data lineage, and how would you trace a wrong number in a management report back to its source?

   What a good answer covers

   • Defines lineage as the path data takes from source to report, including transformations.
   • Starts from the report field and works back through each step.
   • Checks transformation logic, joins and filters, then the source data.
   • Records findings in the catalogue so the fix is visible to others.

8. What governance controls would you ask for before a team deploys a machine learning model trained on customer data?

   What a good answer covers

   • Checks that the training data was collected and used for a purpose customers were told about.
   • Asks for documentation of data sources, model purpose, limits and owner.
   • Looks at bias, explainability and human review for decisions that affect people.
   • Sets monitoring for drift and a process to report incidents.
   • May refer to Singapore's Model AI Governance Framework as a reference.

1. Finance and sales each report a different revenue figure for last quarter. What do you do?

   What a good answer covers

   • Gets both definitions and queries and compares them side by side.
   • Finds whether the gap comes from definitions, timing, filters or data errors.
   • Brings both owners together to agree one definition.
   • Records the agreed definition in the glossary and names the official source.

2. A marketing team wants to upload a customer list to a third-party email tool. How do you respond?

   What a good answer covers

   • Asks what data is included, for what purpose and under what consent.
   • Checks whether customers agreed to marketing and whether any have opted out.
   • Reviews the vendor's data protection terms and where data will be held.
   • Suggests sharing only the fields needed, and records the decision.

3. A key data quality check fails the night before month-end reporting. What are your first steps?

   What a good answer covers

   • Assesses the size of the failure and which reports it affects.
   • Alerts the data owner and engineering team straight away.
   • Agrees whether to delay, report with a caveat or use a corrected extract.
   • Logs the root cause and adds a preventive check afterwards.

1. Tell us about a time you got a reluctant team to follow a data standard.

   What a good answer covers

   • Explains the reason for the standard in terms of the team's own work.
   • Listens to objections and adjusts where the concern was fair.
   • Makes the standard easy to follow with templates or automation.
   • Shows a result, such as fewer errors or less rework.

2. Describe a time you had to explain a data risk to a senior manager who was not technical.

   What a good answer covers

   • Uses plain words and a concrete example instead of jargon.
   • States the impact on customers, cost or compliance.
   • Gives clear options and a recommendation.
   • Follows up on the decision.

3. Tell us about a mistake you made with data and what you changed afterwards.

   What a good answer covers

   • Owns the mistake without blaming others.
   • Explains how it was found and how quickly it was reported.
   • Describes the fix and the new check or process added.
   • Shows what the experience taught about checking work.

1. How do the PDPA's rules on purpose, retention and protection shape a retention schedule for customer data?

   What a good answer covers

   • Says organisations should keep personal data only as long as there is a legal or business need.
   • Links each dataset to a purpose and a retention period agreed with legal.
   • Plans secure deletion or anonymisation when the period ends.
   • Notes that other laws, such as tax record rules from the Inland Revenue Authority of Singapore (IRAS), may require longer retention.

Common questions about a data governance specialist job

Sources