A medical company needs to store clinical data. The data includes personally identifiable information (PII) and protected health information (PHI). An ML engineer needs to implement a solution to ensu

Question

Options

• AStore the clinical data in Amazon S3 buckets. Use AWS Glue DataBrew to mask the PII and PHI
• BUpload the clinical data to an Amazon Redshift database. Use built-in SQL stored procedures to
• CUse Amazon Comprehend to detect and mask the PII before the data is used for model training.
• DCreate an AWS Lambda function to encrypt the PII and PHI. Program the Lambda function to

• A
  3% (1)
• B
  3% (1)
• C
  83% (25)
• D
  10% (3)

Explanation

Option C is correct because Amazon Comprehend is a purpose-built NLP service that can detect and redact PII entities (names, SSNs, diagnoses, etc.) directly in text data before it ever reaches a training pipeline - making it the most targeted, ML-native solution for this use case.

Why the distractors fall short:

• A (Glue DataBrew): DataBrew is a data preparation tool for structured/tabular data; it can mask values, but it's not designed to detect unstructured PII/PHI in clinical text and doesn't integrate cleanly into an ML training pipeline.
• B (Redshift stored procedures): Redshift is an OLAP data warehouse, not a PII detection system - stored procedures can transform data but cannot identify which fields contain PII/PHI automatically.
• D (Lambda + encryption): Encryption does not remove or mask PII/PHI; encrypted data is still sensitive. Encrypting before training doesn't prevent the model from learning on that data if decryption happens upstream.

Memory tip: Think "Comprehend = Comprehend what's in the text." If the goal is to find and neutralize sensitive information in unstructured clinical text before ML training, Comprehend is the AWS service that does the NLP heavy lifting - detect first, mask second, train safely third.

No community discussion yet for this question.