Pass with professional NCP-ADS actual quiz materials

You are designing an accelerated ETL pipeline to process large-scale datasets in a data science workflow.
Which of the following are key considerations when selecting the right tools and methods for implementing this pipeline? (Select two)

• A. Ensuring the ETL pipeline uses only batch processing for data ingestion.
• B. Using GPU-accelerated libraries such as RAPIDS for data transformation to enhance processing speed.
• C. Using a single storage location for both raw and transformed data to simplify the workflow.
• D. Leveraging parallel processing and distributed computing frameworks like Apache Spark to speed up the transformation phase.
• E. Relying on traditional single-threaded processing for the extraction phase to reduce complexity.

A machine learning engineer is tasked with deploying a real-time image classification model as part of an MLOps pipeline. The model requires low-latency inference and must handle high-throughput requests efficiently.
Which of the following deployment strategies is the most suitable for leveraging GPU acceleration?

• A. Deploying the model using NVIDIA Triton Inference Server with TensorRT optimizations.
• B. Running the model inference on a multi-core CPU server with batch processing enabled.
• C. Using a REST API wrapper to load the model dynamically into memory for each request.
• D. Using an Apache Spark cluster with distributed CPU-based inference.

You are working with a cuDF DataFrame and need to convert a column named sales from float64 to int32 to save memory.
Which of the following is the correct and most efficient way to perform this conversion in cuDF?

• A. df['sales'].convert_dtypes('int32')
• B. df['sales'] = df['sales'].astype('int32')
• C. df['sales'] = df['sales'].to_numeric('int32')
• D. df['sales'].apply(lambda x: int(x))

A data science team is deploying a deep learning model for real-time inference. The model is optimized for inference on an NVIDIA A100 GPU, but the team notices that inference latency is higher than expected.
Which of the following optimizations is most effective in reducing inference latency?

• A. Enable CPU offloading to balance the workload between the CPU and GPU.
• B. Increase the batch size significantly to improve GPU utilization.
• C. Use mixed-precision inference with TensorRT to accelerate computation.
• D. Reduce the model size by randomly pruning neurons without retraining.

A data scientist needs to process a dataset containing 10 million records, performing transformations and exploratory data analysis (EDA). The processing needs to be efficient but does not require high- performance multi-GPU execution.
Which of the following libraries provides the best balance between usability and performance?

• A. Dask DataFrame, since it automatically parallelizes computations even when the dataset fits in memory.
• B. Pandas, as it provides a simple API and works well for datasets that fit within system memory.
• C. cuDF, since GPU acceleration will still provide a speedup even for moderately sized datasets.
• D. Spark DataFrame, as it is optimized for distributed processing and scales well even for 10 million records.