NVIDIA NCA-GENM: NVIDIA-Certified Associate: Generative AI Multimodal

Last checked October 22, 2026 against the official NVIDIA exam page.

1. Question 1Deployment and optimization

   A multimodal model must serve many concurrent requests on NVIDIA GPUs with high throughput. Which NVIDIA tool is designed for model serving and dynamic batching?

   1. ANVIDIA Triton Inference Server
   2. BNVIDIA Omniverse Enterprise Suite
   3. CCUDA C++ compiler toolchain alone
   4. DNVIDIA DALI image decoding library
   Show answer and explanationHide answer

   Correct answer: A

   Triton serves models across frameworks, with features such as dynamic batching and concurrent model execution for high throughput. DALI accelerates data preprocessing, Omniverse is for 3D simulation, and the CUDA compiler does not serve models.

2. Question 2Multimodal foundations

   A system must answer questions about photos using both the image and a text question. Which model type is designed to jointly process these two modalities?

   1. AA tabular regression model
   2. BA k-means clustering model
   3. CA vision-language model
   4. DA text-only language model with no image encoder
   Show answer and explanationHide answer

   Correct answer: C

   Vision-language models encode images and text into a shared or aligned representation so they can reason across both. Text-only models lack an image encoder, and regression or clustering models do not process images with language.

3. Question 3Generative model architectures

   Which generative approach creates images by gradually removing noise from a random starting point over many steps?

   1. ADiffusion model
   2. BDecision tree ensemble
   3. CPrincipal component analysis
   4. DSupport vector machine
   Show answer and explanationHide answer

   Correct answer: A

   Diffusion models learn to reverse a gradual noising process, producing images by iterative denoising. Decision trees, PCA, and SVMs are not generative image models of this kind.

4. Question 4Data preparation and evaluation

   A multimodal dataset pairs images with captions, and some captions do not describe their images. What is the most important data quality step?

   1. ATrain anyway because the model will ignore bad pairs
   2. BFilter or correct mismatched image-caption pairs before training
   3. CRemove the captions from every affected pair to avoid mismatch
   4. DConvert every image to black and white before training begins
   Show answer and explanationHide answer

   Correct answer: B

   Mismatched pairs teach the model incorrect associations, so filtering or correcting them improves alignment. Training anyway, removing captions, and colour changes do not address the mismatch.

5. Question 5Deployment and optimization

   Which NVIDIA library optimizes a trained model for low-latency inference by applying graph optimizations and reduced-precision kernels?

   1. ANVIDIA Riva voice samples only
   2. BcuDF
   3. CRAPIDS Dask
   4. DTensorRT
   Show answer and explanationHide answer

   Correct answer: D

   TensorRT builds optimized inference engines through layer fusion, kernel selection, and reduced-precision execution. cuDF and RAPIDS handle dataframes and analytics, and Riva sample projects are not an inference optimizer.