An external consultant will develop a complex Dataflow transform on a project containing private user data. How do you protect user privacy while enabling development?

Creating an anonymized sample dataset in a separate project is the best way to protect user privacy while allowing the consultant to develop the Dataflow transform. This approach ensures the consultant never accesses sensitive production data, fulfilling privacy requirements. Granting the consultant the Project Viewer role would allow them to see the private data, violating privacy. Granting the Cloud Dataflow Developer role would also give them access to the project's resources, potentially including sensitive data, and doesn't address the core privacy concern. Providing a service account still grants access to the project and its data, which is not ideal for privacy with an external consultant.