| Topic | Details |
| Topic 1 | - Business Strategies for a Successful Generative AI Solution: This section of the exam measures the skills of Cloud Architects and evaluates the ability to design, implement, and manage enterprise-level generative AI solutions. It covers the decision-making process for selecting the right solution, integrating AI into an organization, and measuring business impact. A strong emphasis is placed on secure AI practices, highlighting Google¡¯s Secure AI Framework and cloud security tools, as well as the importance of responsible AI, including fairness, transparency, privacy, and accountability.
|
| Topic 2 | - Google Cloud¡¯s Generative AI Offerings: This section of the exam measures the skills of Cloud Architects and highlights Google Cloud¡¯s strengths in generative AI. It emphasizes Google¡¯s AI-first approach, enterprise-ready platform, and open ecosystem. Candidates will learn about Google¡¯s AI infrastructure, including TPUs, GPUs, and data centers, and how the platform provides secure, scalable, and privacy-conscious solutions. The section also explores prebuilt AI tools such as Gemini, Workspace integrations, and Agentspace, while demonstrating how these offerings enhance customer experience and empower developers to build with Vertex AI, RAG capabilities, and agent tooling.
|
| Topic 3 | - Fundamentals of Generative AI: This section of the exam measures the skills of AI Engineers and focuses on the foundational concepts of generative AI. It covers the basics of artificial intelligence, natural language processing, machine learning approaches, and the role of foundation models. Candidates are expected to understand the machine learning lifecycle, data quality, and the use of structured and unstructured data. The section also evaluates knowledge of business use cases such as text, image, code, and video generation, along with the ability to identify when and how to select the right model for specific organizational needs.
|
| Topic 4 | - Techniques to Improve Generative AI Model Output: This section of the exam measures the skills of AI Engineers and focuses on improving model reliability and performance. It introduces best practices to address common foundation model limitations such as bias, hallucinations, and data dependency, using methods like retrieval-augmented generation, prompt engineering, and human-in-the-loop systems. Candidates are also tested on different prompting techniques, grounding approaches, and the ability to configure model settings such as temperature and token count to optimize results.
|