IBM AI Engineering Practice Test

Answer choices

1. A. Tracking supports governance, lifecycle management, and accountability
2. B. Metadata tracking is used only for browser bookmarks
3. C. Tracking forces every deployment back into development mode
4. D. Metadata can replace evaluation results entirely

Wrong-answer review

• B. Metadata tracking is used only for browser bookmarks: Metadata tracking is used only for browser changes the condition being evaluated in this item.
• C. Tracking forces every deployment back into development mode: Tracking forces every deployment back into development addresses nearby material but misses the Governance point.
• D. Metadata can replace evaluation results entirely: Metadata can replace evaluation results entirely may sound plausible, but it does not answer this item.

Answer choices

1. A. Delete the project after training
2. B. Deploy the asset from a deployment space
3. C. Recreate the training data as a prompt template
4. D. Wait for a new model before considering deployment

Answer choices

1. A. That the project uses only PDF files
2. B. That the deployment space has no description text
3. C. That the project has an associated watsonx.ai Runtime service
4. D. That the vector index uses no embeddings

Answer choices

1. A. Export the scikit-learn model as a prompt template using Prompt Lab
2. B. Save the model as a raw CSV file containing only the model weights and load it manually into AutoAI
3. C. Use `client.repository.store_model` with the scikit-learn model object, specifying its name and a compatible software specification
4. D. Custom models must be translated into Spark SQL queries before deployment in watsonx

Answer choices

1. A. A multilingual text embedding model
2. B. A monolingual English embedding model with automatic rule-based dictionary lookup
3. C. A pure visual embedding model trained on screenshots
4. D. An un-tokenized character hash generator

Answer choices

1. A. Sampling decoding with a Temperature of 1.0
2. B. Greedy decoding with a high Repetition Penalty
3. C. Greedy decoding with a Temperature of 0.0
4. D. Sampling decoding with a low Top-P value

Answer choices

1. A. Convert the DataFrame to a Pandas DataFrame using `.toPandas()` and apply a lambda function
2. B. Iterate through the Spark DataFrame using a Python `for` loop and store the results in a local list
3. C. Save each row to a local text file, process them in parallel with a shell script, and reload them
4. D. Use the built-in Spark SQL `concat` or `concat_ws` functions inside a `select` or `withColumn` transformation

Answer choices

1. A. Only a project rename and a browser refresh
2. B. A vector index for grounding plus prompt-template evaluation and tracking
3. C. AutoAI with no source documents and no evaluation plan
4. D. A deployment space with no model, no prompt, and no test data

Answer choices

1. A. project_id or space_id
2. B. tenant_credential_token
3. C. billing_account_arn
4. D. deployment_target_uri

Answer choices

1. A. Map each prompt variable to the correct test-data column
2. B. Delete all variables from the template
3. C. Convert the project into a deployment space first
4. D. Disable reference outputs in the evaluation set