NVIDIA-Certified Associate: Generative AI and LLMs

Published weight 30% (largest area). ML and NLP fundamentals, neural network and transformer architecture, LLM foundations, prompt engineering, and GenAI application patterns such as RAG and chatbots.

• Explain core machine learning and neural network concepts, including transformer architecture and how it powers modern LLMs
• Apply prompt engineering techniques and understand differences among large language models
• Describe generative AI application patterns such as retrieval-augmented generation (RAG) and conversational agents

Published weight 24%. Python and software development for LLM applications, common libraries (spaCy, NumPy, Keras), LLM integration and deployment, and evaluating scalability and system integration.

• Use Python libraries and frameworks to build and integrate LLM-powered applications
• Deploy and integrate LLMs into software systems, evaluating scalability and performance
• Apply data preprocessing and feature engineering within an application development workflow

Published weight 22%. Designing, running, and evaluating experiments including A/B testing, zero-shot and few-shot testing, prompt evaluation, and interpreting results.

• Design and run experiments such as A/B tests and zero-shot / few-shot evaluations
• Evaluate prompts and model outputs, and interpret experimental results to guide iteration

Published weight 14%. Data cleaning and preparation, data visualization, and statistical metrics such as loss functions and explained variance ratio used to assess models and datasets.

• Clean, prepare, and visualize data for machine learning and LLM workflows
• Interpret statistical metrics such as loss functions and explained variance ratio

Published weight 10%. AI ethics, alignment, bias identification and minimization, and balancing privacy, security, and user consent in generative AI systems.

• Identify and minimize bias and apply responsible, ethical AI practices
• Balance data privacy, security, and user consent when building AI applications