Google Cloud Generative AI Leader Certification Notes

Notes for my Cloud certifications.

Google Cloud Generative AI Leader Certification Notes

These are my personal notes for the Google Cloud Generative AI Leader Certification, taken during following the Cloud Skills Boost Generative AI Leader path.

Overview:

Helpful resources:

1. Data and Machine Learning Fundamentals

Data as the Foundation of AI

Data is the foundation of any AI system. Data quality and accessibility are essential for effective AI development.
Data can be structured or unstructured, each requiring different analysis techniques.

Key dimensions of data quality:

Understanding the types and quality of your data is crucial for successful AI initiatives.


Machine Learning Approaches

Machine learning models can be trained using:

The choice of approach depends on the specific task and the nature of the data available.


The ML Lifecycle

The ML lifecycle encompasses several key stages:

Google Cloud provides a comprehensive suite of tools to support each stage of this lifecycle.

Vertex AI helps with model training and deployment, while various data tools support ingestion, preparation, and management.

By understanding and effectively managing this lifecycle, organizations can maximize the value of their initiatives and ensure long-term success.


2. Model Development with Vertex AI

Model Training

The process of creating your ML model using data is called model training.

Vertex AI provides:


Model Deployment

Model deployment is the process of making a trained model available for use.

Vertex AI simplifies this with:


Model Management

Managing and maintaining your models over time is critical.

Google Cloud offers:


3. Foundation Models and Generative AI

Deep learning provides the core technology.
Foundation models are powerful architectures built on deep learning.
Generative AI is the application of these models to create new, original content.


Vertex AI for Generative AI

Vertex AI streamlines integration of advanced AI capabilities into business applications:

These models empower businesses to enhance customer experiences, increase productivity, foster innovation, and improve decision-making.


Google-Developed Models on Vertex AI

Gemini is designed to handle multiple data types, while Gemma is optimized for lighter, specialized deployments.


Considerations for Choosing Generative AI Models

Google Cloud offers a suite of foundation models with unique strengths and capabilities.


4. Limitations of Foundation Models

Common Limitations


5. Techniques to Overcome Limitations

Grounding

Connect the AI’s output to verifiable sources—like giving AI a reality check.

Benefits:


Retrieval-Augmented Generation (RAG)

RAG grounds outputs in real, verifiable sources, improving accuracy and relevance.


Prompt Engineering

The most rapid, straightforward approach to guide models.


Fine-Tuning

When prompting isn’t enough, fine-tuning adapts a model to specific needs.

Use Cases:

Vertex AI provides tooling to facilitate tuning.


6. Humans in the Loop (HITL)

Even the best models benefit from human oversight.

Key use cases:


7. Secure AI

Preventing intentional harm to AI applications.

Key risks:

Google Cloud’s SAIF framework provides tools to help build and maintain secure AI systems.


8. Responsible AI

Ensuring AI avoids both intentional and unintentional harm.


Transparency

Users need to know how their information is used and how AI systems work.


Privacy

Protecting privacy often involves anonymization or pseudonymization.


Data Quality, Bias, and Fairness

High-quality data is essential for ethical AI.

Example: A resume-screening tool favoring certain demographics due to biased training data.


Accountability and Explainability

Fairness requires accountability.

Vertex Explainable AI helps:


AI development is governed by evolving legal frameworks.

Key considerations:

Legal compliance is essential for building trustworthy AI systems.


9. Agents and Gen AI Applications

What Can Agents Do?

Gen AI agents process information, reason over complex concepts, and take action.

Applications include:


Defining a Gen AI Agent

An application that observes the world and acts on it using its tools to achieve goals.

Capabilities:


Agent Workflows

Conversational Agents


Workflow Agents


Advanced Prompt Engineering Frameworks

Examples include ReAct and Chain-of-Thought (CoT).


10. Vertex AI MLOps Tools

Manage the ML lifecycle with built-in tools.


11. Building Models with Vertex AI

Two main options:


12. Gemini Nano

Google’s most efficient, compact AI model for edge deployment.

Tools: Lite Runtime (LiteRT), Gemini Nano


13. Gemini for Google Workspace

Access Gemini’s generative AI features within Gmail, Docs, Sheets, Meet, and Slides.


14. Prompting Techniques


Role Prompting

Guide the model by assigning a persona.

Examples:


Prompt Chaining

Create complex interactions where each prompt builds on the last.


Grounding

Ensures outputs are based on verifiable, specific sources.


Retrieval-Augmented Generation (RAG)


15. NotebookLM

An AI-first notebook grounded in your own documents.

Capabilities:

Plus: Adds capacity, customization, usage analytics.
Enterprise: Extra privacy, compliance, IAM controls.

Learn more


16. Sampling Parameters and Settings


17. Google AI Studio vs. Vertex AI Studio

Feature Google AI Studio Vertex AI Studio
Audience Experimenters, early-stage users Developers building production systems
Features Easy Gemini API access Advanced tools for the ML lifecycle

18. Prompt Engineering Techniques

ReAct Framework

Combines reasoning and action.

Steps:

Benefits:


Chain-of-Thought (CoT) Prompting

Guides the model through step-by-step reasoning.

Benefits:

Techniques:


19. Reasoning Loop with Tools

ReAct Cycle:

  1. Reasoning (Tool Selection)

  2. Acting (Tool Execution)

  3. Observation

  4. Iteration


20. How RAG Works with Tools


21. Conversational Agents and Playbooks

Define step-by-step behaviors using linked external tools and data stores.


22. Metaprompting

Enables dynamic, adaptable prompt creation and interpretation.


23. Agentspace

Centralized platform to manage AI agents using company data.

Agentspace vs. NotebookLM

Feature NotebookLM Agentspace
Purpose Deep dive into specific documents Enterprise AI assistant across systems
Scope Only user-provided sources All connected business systems
Integration Can connect with NotebookLM Enterprise Unified search and automation

Additional Helpful Resources: