DP-900 · 1 Describe core data concepts
Data engineer role
Exam objective: Describe responsibilities for data engineers
A data engineer builds and monitors the pipelines that integrate and prepare data across an organization, working across many kinds of data stores.
A data engineer collaborates with stakeholders to design and implement the data-related workloads an organization needs: pipelines that ingest data, routines that clean it, transformations that reshape it, and the stores that hold it for analytical use.
That work spans a wide range of platform technologies rather than a single one. A data engineer might work with relational databases, nonrelational databases, file stores and data streams in the same solution, choosing whichever fits the source or destination of the data.
Once a pipeline is built, the data engineer owns monitoring it, making sure data loads keep performing as expected over time. They are also accountable for the privacy of the data as it travels from on-premises systems into the cloud, and for applying the organization's governance rules along the way.
Microsoft Learn notes that AI tools can help generate transformation code or suggest configurations in natural language, but the architectural decisions and data quality judgment remain the data engineer's own.
On the exam, look for pipelines, ingestion, data cleansing and transformation, or moving data between systems.
Key points
- A data engineer designs and implements data workloads, including ingestion pipelines and the cleansing and transformation activities that run inside them.
- The role spans many technologies: relational databases, nonrelational databases, file stores and data streams, not just one of them.
- A data engineer owns monitoring the pipelines they build, to confirm that data loads keep performing as expected.
- They are also responsible for the privacy of data as it moves from on-premises systems into the cloud.
Exam trap
Turning the data into a report or a chart is the data analyst's job. The data engineer's responsibility ends at delivering clean, reliable data to the store the analyst queries.
Check yourself
A pipeline that moves data from an on-premises relational database into a cloud analytical store starts failing intermittently, and loads stop performing as expected. Whose responsibility is monitoring and fixing this?
Go deeper on Microsoft Learn
Checked against Microsoft Learn on October 1, 2026.