AWS Certified AI Practitioner
Secure Data Engineering for AI
The four practices the exam names for the data pipeline behind a model: assessing data quality, privacy-enhancing technologies, data access control, and data integrity.
Intermediate 22 minutes 5 Learning Objectives
- Explain why data problems become permanent once a model is trained on them
- Use AWS Glue Data Quality rulesets to assess a dataset before it reaches a model
- Compare privacy-enhancing technologies and choose one for a given constraint
- Design layered data access control across S3, IAM, and Lake Formation
- Describe the integrity controls that make a training corpus reproducible and tamper-evident