
Ultimate access to all questions.
Your team is tasked with designing a comprehensive fraud detection system for a major international bank. The system must efficiently process real-time transactions from various banking applications, standardize the data format, store the data in real-time with necessary statistical aggregations, periodically train a machine learning model for accurate outlier detection, and provide fraud probabilities for each transaction. The solution should prioritize minimizing manual labeling efforts and reducing software development overhead. Given these requirements, which three Google Cloud products would you choose to build the most effective and efficient solution? (Choose three options)
A
Cloud Dataproc for processing and analyzing large datasets with Apache Spark and Hadoop
B
BigQuery for real-time analytics and data warehousing with built-in machine learning capabilities
C
Cloud Dataflow for unified stream and batch data processing with minimal latency
D
Cloud Dataprep for data preparation and cleaning without the need for coding
E
Cloud Composer for workflow orchestration managed with Apache Airflow
F
Cloud Pub/Sub for scalable and reliable real-time messaging