PMLE practice questions 361–364

1. Question 361 · 364 in bank

   You work for a media company. You are developing a solution that requires transcription and diarizing (speaker identification) of live audio streams in near real-time. You are using an optimized WhisperX model built on the PyTorch framework that uses several libraries and weights sourced from the Hugging Face Hub. The deployment must include proprietary Python post-processing logic including speaker diarization and alignment correction that executes immediately after model inference. You need your solution to be efficient, scalable, handle high throughput, and minimize serving latency. How should you deploy this model and its required logic?

   • ADeploy the WhisperX model artifacts to an Agent Platform endpoint using a pre-built PyTorch serving container configured with a GPU. Host the custom post-processing scripts in a Cloud Run function that receives the endpoint’s prediction output and performs the final diarization.
   • BUse a self-managed Google Kubernetes Engine (GKE) cluster for deployment. Package the model and logic into a container, and configure the cluster’s horizontal pod autoscaler (HPA) to manage throughput.
   • CPackage the WhisperX model artifacts, the specific Python post-processing scripts, and the necessary dependencies (including Hugging Face libraries) into a single, custom serving container image. Deploy this custom image to a GPU-enabled Agent Platform endpoint.
   • DLocate the WhisperX model in the Model Garden and use the “1-click deploy” feature to deploy it to an Agent Platform endpoint. Route the output to a Dataflow pipeline that uses a custom DoFn function to handle the post-processing and logging.

   Show answer

   Answer: C

   Next question

2. Question 362 · 364 in bank

   You deployed a conversational application that uses Gemini. You recently collected thousands of rows of user feedback about the verbosity and accuracy of the model’s responses. The feedback indicated that the responses are factually correct but users want different levels of verbosity depending on the type of question. You need to adjust the model to return responses that are more consistent with users’ expectations. You want to use a scalable solution and minimize application maintenance. What should you do?

   • AExperiment with other models in Model Garden. Perform A/B testing by setting each model as your application’s default model. Choose a model based on the results.
   • BAsk users to indicate all scenarios where they expect concise responses versus verbose responses. Modify the application’s prompt to include these scenarios and their respective verbosity levels. Re-evaluate the verbosity of responses with updated prompts.
   • CImplement a keyword-based routing layer. If the user’s input contains the words “detailed” or “description,” return a verbose response. If the user’s input contains the word “fact,” re-prompt Gemini to summarize the response and return a concise response.
   • DAsk users to provide examples of responses with the appropriate verbosity as a list of question and answer pairs. Use this dataset to perform supervised fine-tuning (SFT) in Agent Platform. Re-evaluate the verbosity of responses with the tuned version.

   Show answer

   Answer: D

   Next question

3. Question 363 · 364 in bank

   You manage an ML workflow for a fraud detection model. The source code for data processing and model training is stored in a version control repository. You need to implement a continuous training (CT) pipeline that automatically runs unit tests and launches a new training run on Agent Platform Pipelines when code is updated in the repository. You want to use a fully managed service to orchestrate this process. What should you do?

   • ACreate a Cloud Run function that is triggered by repository commits. The function runs the unit tests, launches a custom training job, and uses Model Monitoring to validate the new model’s performance before deployment.
   • BDefine a Cloud Workflow to monitor the repository for changes. Configure the workflow on Agent Platform Pipelines to orchestrate and run the unit tests. Use the model evaluation component to validate the new model’s performance before deployment.
   • CConfigure Agent Platform Pipelines to poll the repository for updates. Add a conditional step in the pipeline that uses Model Monitoring to validate the new model’s performance before deployment.
   • DCreate a Cloud Build trigger associated with the repository. Define a build configuration file that runs the unit tests and triggers a job on Agent Platform Pipelines. Use the model evaluation component to validate the model performance before deployment.

   Show answer

   Answer: D

   Next question

4. Question 364 · 364 in bank

   You are deploying a generative AI chatbot for an online gaming community using Gemini Enterprise Agent Platform. During beta testing, you discover that users are able to input malicious prompts that cause the model to generate hate speech and violent content. You need to implement a guardrail to block this content in production with minimal development overhead. What should you do?

   • AUpdate the system instruction (system prompt) of the model to explicitly state: “You are a helpful assistant. Do not generate hate speech or violent content.”
   • BImplement and configure Model Armor to sanitize responses if hateful content is detected.
   • CCreate a supervised fine-tuning (SFT) job in Agent Platform, and tune the model using a curated dataset of safe, policy-compliant gaming conversations.
   • DIntegrate the Cloud Natural Language API’s moderation capabilities to analyze and block toxic user prompts before they reach the chatbot model.

   Show answer

   Answer: B

Keyboard:J Kmove between questionsA–Dselect an answerRshow or hide the answer← →change page

PMLE practice, question 361. 364 questions available.