[NEW] SnowPro Specialty: GenAI

Snowflake Gen AI & LLM Functions (40%): Practical implementation of SNOWFLAKE.CORTEX.COMPLETE, vector functions for similarity, and building end-to-end RAG pipelines using Snowpark.

Snowflake Gen AI Governance (22%): Monitoring costs via CORTEX_FUNCTIONS_USAGE_HISTORY, token pricing models, auditing operations, and setting up AI guardrails.

Snowflake Document AI (12%): Extracting structured data from unstructured formats like PDFs and training custom document models for downstream workflows.

B. VECTOR_L2_DISTANCE or VECTOR_COSINE_SIMILARITY

C. CORTEX_FUNCTIONS_USAGE_HISTORY

D. Document AI model training

E. Snowflake Copilot auto-completion

F. SNOWFLAKE.ML.FORECAST

Correct Answer: B

Explanation:

A (Incorrect): This function shortens text but does not handle vector math or similarity.

C (Incorrect): This view is for cost and usage auditing, not for building RAG logic.

D (Incorrect): Document AI is for extraction from PDFs/images, not for the vector retrieval phase of RAG.

E (Incorrect): Copilot is a developer assistant tool, not a runtime function for similarity search.

F (Incorrect): This is a Time Series ML function and is unrelated to Generative AI embeddings.

Question 2: To manage costs effectively, which Snowflake view should an administrator query to track token consumption specifically for Cortex LLM functions?

B. QUERY_HISTORY

C. CORTEX_FUNCTIONS_USAGE_HISTORY

D. ACCESS_HISTORY

E. SERVERLESS_TASK_HISTORY

F. DATA_TRANSFER_HISTORY

Correct Answer: C

Explanation:

A (Incorrect): This tracks general virtual warehouse usage, but Cortex often runs on serverless compute that requires this specific AI usage view.

B (Incorrect): While it shows the query, it doesn't provide the granular token-level breakdown needed for AI cost optimization.

D, E, F (Incorrect): These monitor data access, tasks, and cloud egress, which are not the primary metrics for LLM token pricing.

Question 3: When using Document AI, what is the purpose of the "Value Inspection" and "Training" phase within the Snowflake UI?

B. To confirm the model is correctly extracting fields and provide corrections to improve accuracy.

C. To encrypt the PDF files before they reach the LLM.

D. To translate the document from English to French.

E. To delete the original unstructured files from the stage.

F. To resize the images for better web viewing.

Correct Answer: B

Explanation:

A (Incorrect): This phase is about the model's extraction logic, not writing general SQL.

C (Incorrect): Security and encryption are handled at the stage level, not during the model training phase.

D (Incorrect): While Cortex can translate, the Document AI training phase is focused on structured data extraction.

E, F (Incorrect): These are file management and preprocessing tasks, not model training activities.

You can retake the exams as many times as you want

This is a huge original question bank

You get support from instructors if you have questions

Each question has a detailed explanation

Mobile-compatible with the Udemy app

30-days money-back guarantee if you're not satisfied