GES-P02 Sample Questions & Answers
Cortex LLM functions, retrieval, agents and prompt engineering carry the most weight, next to access control and cost governance for generative AI, document parsing pipelines, and the principles behind open-source and custom models on the platform.
Launch the full GES-P02 simulator →Showing 6 of 12 free samples.
- Question 1Beginner
Snowflake for Gen AI Overview · Snowpark Container Services and Snowflake Model Registry
True or False: Snowpark Container Services requires organizations to provision external AWS or Azure GPU virtual machines manually before deploying open-source Large Language Models in Snowflake.
Show answer & explanation
Correct answer: B
False. Snowpark Container Services provides fully managed compute pools, including GPU-backed instances, directly within Snowflake. Customers do not need to manually provision or manage underlying cloud provider VMs; they simply create a Snowflake compute pool specifying the desired instance family.
- Question 2Advanced
Snowflake for Gen AI Overview · Snowflake Cortex AI overview and the Data Cloud approach to Gen AI
A retail enterprise is architecting a new application to categorize millions of customer reviews into positive, neutral, or negative sentiments. The engineering team is debating between using the fully managed
SNOWFLAKE.CORTEX.SENTIMENTfunction versus fine-tuning an open-source model and deploying it via Snowpark Container Services (SPCS).They have strict requirements to minimize infrastructure maintenance overhead and want the pricing to scale exactly with the number of tokens processed, without paying for idle compute time.
Which architectural approach BEST meets these specific constraints?
Show answer & explanation
Correct answer: D
Cortex LLM functions are fully managed and serverless. They abstract away infrastructure management and bill based on the actual tokens processed. Deploying a custom model via SPCS requires provisioning compute pools, which incur costs as long as they are running (idle time is billed), and introduces infrastructure maintenance overhead.
- Question 3Intermediate
Snowflake for Gen AI Overview · Gen AI use cases on Snowflake
A developer wants to accelerate their SQL and Python development within the Snowflake UI using an AI assistant that provides auto-completion, code generation, and debugging help. Which Snowflake capability is designed specifically for this use case?
Show answer & explanation
Correct answer: A
Cortex Code is Snowflake's AI-powered development assistant that helps developers write SQL and Python faster by offering code generation, auto-completion, and debugging capabilities directly within Snowflake interfaces.
- Question 4Beginner
Snowflake for Gen AI Overview · Snowflake Cortex AI overview and the Data Cloud approach to Gen AI
The deployment model for Snowflake Cortex LLM functions is considered __________ because users do not need to provision GPUs, manage infrastructure, or maintain model weights.
Show answer & explanation
Correct answer: A
Snowflake Cortex Gen AI capabilities are provided as serverless functions. This means Snowflake handles all the underlying compute, GPU provisioning, and model scaling automatically behind the scenes.
- Question 5Beginner
Snowflake Gen AI Functions · Task-specific and general LLM functions
A data analyst needs to generate a short, concise overview of long customer support transcripts. They can use either
SNOWFLAKE.CORTEX.SUMMARIZEorSNOWFLAKE.CORTEX.COMPLETEwith a prompt asking for a summary. Why isSUMMARIZEgenerally the preferred best practice for this specific task?Show answer & explanation
Correct answer: C
Task-specific functions like SUMMARIZE, TRANSLATE, and SENTIMENT use smaller, highly optimized models under the hood. They are more cost-effective and have lower latency than using the general-purpose COMPLETE function with a large model for standard tasks.
- Question 6Intermediate
Snowflake Gen AI Functions · AISQL and the expanded AI function suite
A data engineer is preparing a table of clinical trial notes for broader internal analysis. The notes contain Personally Identifiable Information (PII) such as patient names and phone numbers. Which AISQL function should the engineer apply to automatically obscure this sensitive information before granting access to downstream users?
Show answer & explanation
Correct answer: A
AI_REDACT is part of the expanded AISQL function suite. It identifies and masks Personally Identifiable Information (PII) and other sensitive entities within unstructured text fields directly in the SQL query.
graph LR A[Raw Note: 'Call John at 555-0199'] --> B(AI_REDACT) B --> C[Output: 'Call [NAME] at [PHONE]']
Ready for the real thing?
The full GES-P02 simulator has every exam-style question, timed mode, and instant scoring.