Snowflake SnowPro® Specialty: Gen AI Certification Sample Questions:

1. A developer is building a client application that interacts with a Snowflake Cortex Agent using its REST API. They are implementing multi- turn conversation support. Which of the following is the most critical aspect for maintaining conversational context over multiple API calls?

A) Only the last user prompt and the immediately preceding agent response should be sent to conserve token usage.
B) The agent internally summarizes long conversations, and the client receives only a condensed summary_token to pass for subsequent turns.
C) A specific session_id parameter, generated at the start of the conversation, must be sent with each request to link turns.
D) The REST API automatically identifies and reuses context from previous requests based on the Authorization token, eliminating the need to pass full history.
E) The client application must include all prior user prompts and assistant/analyst responses in a messages array in each new API request.

2. A data engineering manager needs to audit Cortex LLM function costs to identify specific SQL queries that are unexpectedly high in token consumption for the 'llama3.1-8b' model. They require granular analysis of prompt, completion, and guardrail token usage for these queries. Which of the following Snowflake methods or views would provide the necessary insights?

A) Option E
B) Option A
C) Option B
D) Option D
E) Option C

3. A data engineering team needs to configure their Snowflake environment to process documents using AI_PARSE_DOCUMENT and generate text embeddings using EMBED_TEXT_1024 with the voyage-multilingual-2 model. Their Snowflake account is in a region where these specific capabilities or models are only available via cross-region inference. The team needs to ensure these functions work correctly without constant region-specific model selection. Which of the following is the correct configuration action and an important consideration?

A) Option E
B) Option A
C) Option B
D) Option D
E) Option C

4. A financial institution wants to develop a Snowflake-based pipeline to process call transcripts from their customer support. The pipeline needs to perform two main tasks: first, ''summarize very lengthy technical support calls'' (up to 20,000 tokens per transcript) into concise actionable insights, and second, ''classify the sentiment'' of these calls as 'positive', 'neutral', or 'negative'. Given these requirements for integration into SQL data pipelines, which combination of Snowflake Cortex functions and prompt engineering considerations would be most appropriate?

A) Option E
B) Option A
C) Option B
D) Option D
E) Option C

5. A data platform architect is evaluating the integration of 'SNOWFLAKE.CORTEX.TRANSLATE into several automated data pipelines. One pipeline involves real-time translation of messages for a chat application, while another is for batch processing of archived documents. The architect is considering various Snowflake features for orchestration and deployment. Which of the following considerations about 'SNOWFLAKE.CORTEX.TRANSLATE is accurate?

A) The Snowflake managed model used by the 'TRANSLATE' function has a context window of 4,096 tokens, meaning texts longer than this will be truncated before translation.
B) When using 'TRANSLATE within a Snowpark Python User-Defined Function (UDF), the raw text data must be explicitly moved out of Snowflake's network boundary to the underlying LLM service for translation.
C) If 'TRANSLAT' is not natively available in the account's primary Snowflake region, cross-region inference cannot be enabled, thus preventing its use.
D) The 'TRANSLATE' function can be seamlessly integrated into a dynamic table's ' SELECT statement to provide continuous, automated translation with minimal configuration.
E) To manage potential failures in a production pipeline, 'SNOWFLAKCORTEX.TRANSLATE should be wrapped in 'TRY COMPLETE' for robust error handling, returning NULL on failure instead of an error.

Solutions: