Google Professional Machine Learning Engineer Quick Facts (2025)

Exam Domain Breakdown

Domain 1: Section 1: Architecting low-code AI solutions (13% of the exam)

1.1 Developing ML models by using BigQuery ML. Considerations include:

• Building the appropriate BigQuery ML model (e.g., linear and binary classification, regression, time-series, matrix factorization, boosted trees, autoencoders) based on the business problem
• Feature engineering or selection by using BigQuery ML
• Generating predictions by using BigQuery ML

1.1 summary: This section emphasizes how BigQuery ML can accelerate value creation by allowing practitioners to build ML models directly within the data warehouse. Without leaving BigQuery, analysts and engineers can use familiar SQL syntax to design models tailored to business needs, facilitating collaboration and faster deployment cycles. It also highlights the importance of appropriate feature selection, model choice, and creating workflows that support accurate prediction at scale.

By mastering these capabilities, you will better understand how to connect data modeling directly to business outcomes. This ensures that data teams can rapidly prototype and deploy predictive capabilities while benefiting from Google Cloud’s powerful infrastructure, making it easier to scale insights throughout an organization.

1.2 Building AI solutions by using ML APIs or foundational models. Considerations include:

• Building applications by using ML APIs from Model Garden
• Building applications by using industry-specific APIs (e.g., Document AI API, Retail API)
• Implementing retrieval augmented generation (RAG) applications by using Vertex AI Agent Builder

1.2 summary: This section is centered on simplifying the development of AI-driven applications by leveraging pre-built APIs and foundation models. You will learn how Model Garden provides a wide range of options for out-of-the-box solutions and industry-specific APIs that can directly address use cases like document extraction, retail predictions, or conversational intelligence.

By practicing these integrations, you will see how ML APIs and retrieval augmented generation techniques can dramatically reduce development time while ensuring performance and scalability. The focus is on practical strategies that allow teams to deploy sophisticated AI applications without needing to reinvent the core machine learning models.

1.3 Training models by using AutoML. Considerations include:

• Preparing data for AutoML (e.g., feature selection, data labeling, Tabular Workflows on AutoML)
• Using available data (e.g., tabular, text, speech, images, videos) to train custom models
• Using AutoML for tabular data
• Creating forecasting models by using AutoML
• Configuring and debugging trained models

1.3 summary: This section showcases how AutoML empowers teams to create custom machine learning solutions with minimal manual model coding. You will learn techniques to prepare and label data across modalities, configure AutoML workflows, and build specialized models such as forecasting systems. It emphasizes how AutoML accelerates solution creation while maintaining strong predictive performance.

As you progress, you will explore strategies for debugging models and interpreting results for production readiness. This ensures that AutoML becomes not just a productivity tool, but also a structured approach that makes model building more reliable and accessible for organizations of any size.

Domain 2: Collaborating within and across teams to manage data and models (14% of the exam)

2.1 Exploring and preprocessing organization-wide data (e.g., Cloud Storage, BigQuery, Spanner, Cloud SQL, Apache Spark, Apache Hadoop). Considerations include:

• Organizing different types of data (e.g., tabular, text, speech, images, videos) for efficient training
• Managing datasets in Vertex AI
• Data preprocessing (e.g., Dataflow, TensorFlow Extended [TFX], BigQuery)
• Creating and consolidating features in Vertex AI Feature Store
• Privacy implications of data usage and/or collection (e.g., handling sensitive data such as personally identifiable information [PII] and protected health information [PHI])
• Ingesting different data sources (e.g., text documents) into Vertex AI for inference

2.1 summary: This section highlights how managing large-scale organizational data requires not just technical infrastructure but also careful planning. You will learn how to organize multimodal datasets for efficiency, leverage Vertex AI capabilities, and implement preprocessing workflows that ensure consistent model performance.

Additionally, you will gain insight into considerations of data privacy and compliance, a key factor in operational ML. This enables responsible data use while preparing pipelines that are repeatable, scalable, and aligned with organizational policies.

2.2 Model prototyping using Jupyter notebooks. Considerations include:

• Choosing the appropriate Jupyter backend on Google Cloud (e.g., Vertex AI Workbench, Colab Enterprise, notebooks on Dataproc)
• Applying security best practices in Vertex AI Workbench
• Using Spark kernels
• Integrating code source repositories
• Developing models in Vertex AI Workbench by using common frameworks (e.g., TensorFlow, PyTorch, sklearn, Spark, JAX)
• Leveraging a variety of foundational and open-source models in Model Garden

2.2 summary: In this section, the focus is on providing seamless environments for experimentation through Jupyter notebooks. Vertex AI Workbench, Colab Enterprise, and Dataproc ensure that teams have flexible environments suitable for different workloads, whether they require GPU acceleration, Spark support, or tight integration with source control.

You will also deepen your knowledge of integrating common machine learning frameworks within notebooks, while applying strong security practices. By leveraging foundational and open-source models, teams can rapidly test solutions, refine architectures, and collaboratively develop models in line with organizational objectives.

2.3 Tracking and running ML experiments. Considerations include:

• Choosing the appropriate Google Cloud environment for development and experimentation (e.g., Vertex AI Experiments, Kubeflow Pipelines, Vertex AI TensorBoard with TensorFlow and PyTorch) given the framework
• Evaluating generative AI solutions

2.3 summary: This section covers how to track, audit, and optimize ML experiments to accelerate innovation. Vertex AI Experiments, TensorBoard, and Kubeflow Pipelines all provide different options for logging results, comparing metrics, and maintaining a rigorous approach to experimentation.

By understanding these tools, you can manage iterative progress confidently and evaluate results from complex generative AI solutions. This ensures that experimentation not only produces stronger models but also builds knowledge in a repeatable and transparent way across teams.

Domain 3: Scaling prototypes into ML models (18% of the exam)

3.1 Building models. Considerations include:

• Choosing ML framework and model architecture
• Modeling techniques given interpretability requirements

3.1 summary: This section focuses on selecting the correct architecture and frameworks that best match business needs and interpretability goals. By examining the tradeoffs between deep learning, classical algorithms, and interpretable models, you will learn how to design systems that meet organizational requirements.

You will also look at how frameworks such as TensorFlow, PyTorch, or JAX interact with architectural decisions. Mastering these choices ensures your solutions are both technically robust and aligned with end-user trust and explainability needs.

3.2 Training models. Considerations include:

• Organizing training data (e.g., tabular, text, speech, images, videos) on Google Cloud (e.g., Cloud Storage, BigQuery)
• Ingestion of various file types (e.g., CSV, JSON, images, Hadoop, databases) into training
• Training using different SDKs (e.g., Vertex AI custom training, Kubeflow on Google Kubernetes Engine, AutoML, tabular workflows)
• Using distributed training to organize reliable pipelines
• Hyperparameter tuning
• Troubleshooting ML model training failures
• Fine-tuning foundational models (e.g., Vertex AI, Model Garden)

3.2 summary: This section emphasizes building training pipelines that are consistent, distributed, and reliable. You will study the ingestion of diverse file formats, workflows for distributed training, and the application of hyperparameter tuning to improve results.

With these strategies, you will gain expertise in scaling model training for large datasets, as well as troubleshooting complex workloads. Additionally, the section highlights fine-tuning existing foundation models as an efficient way to adapt pre-trained architectures to specific business use cases.

3.3 Choosing appropriate hardware for training. Considerations include:

• Evaluation of compute and accelerator options (e.g., CPU, GPU, TPU, edge devices)
• Distributed training with TPUs and GPUs (e.g., Reduction Server on Vertex AI, Horovod)

3.3 summary: This section highlights how to select hardware that aligns with training cost, performance, and scalability requirements. By comparing CPU, GPU, and TPU usage, you will understand how different accelerators support distinct workloads.

Additionally, the section explores distributed training strategies to ensure optimal performance when scaling across multiple devices. This allows you to balance efficiency, training speed, and cost effectively when deploying resource-intensive ML solutions.

Domain 4: Serving and scaling models (20% of the exam)

4.1 Serving models. Considerations include:

• Batch and online inference (e.g., Vertex AI, Dataflow, BigQuery ML, Dataproc)
• Using different frameworks (e.g., PyTorch, XGBoost) to serve models
• Organizing a model registry
• A/B testing different versions of a model

4.1 summary: This section teaches how to operationalize models by serving them in an efficient, reliable, and scalable manner. You will explore the differences between batch and online inference, the use of Vertex AI and Dataflow, and integration with common frameworks for model deployment.

The section also emphasizes building a strong model registry and running A/B testing to validate performance across versions. These practices boost business confidence by ensuring new models improve outcomes before being widely adopted.

4.2 Scaling online model serving. Considerations include:

• Vertex AI Feature Store
• Vertex AI public and private endpoints
• Choosing appropriate hardware (e.g., CPU, GPU, TPU, edge)
• Scaling the serving backend based on the throughput (e.g., Vertex AI Prediction, containerized serving)
• Tuning ML models for training and serving in production (e.g., simplification techniques, optimizing the ML solution for increased performance, latency, memory, throughput)

4.2 summary: This section guides you through the best practices for scaling models once they are in production. You will examine how to use Vertex AI tools, such as Feature Store and Prediction services, and determine the appropriate scaling strategy for workloads that span public and private endpoints.

You will also consider hardware, model tuning, and techniques for balancing latency and throughput. These lessons ensure that your ML systems can deliver consistent predictions at the required volumes, while remaining efficient and cost-effective.

Domain 5: Automating and orchestrating ML pipelines (22% of the exam)

5.1 Developing end-to-end ML pipelines. Considerations include:

• Data and model validation
• Ensuring consistent data pre-processing between training and serving
• Hosting third-party pipelines on Google Cloud (e.g., MLFlow)
• Identifying components, parameters, triggers, and compute needs (e.g., Cloud Build, Cloud Run)
• Orchestration framework (e.g., Kubeflow Pipelines, Vertex AI Pipelines, Cloud Composer)
• Hybrid or multicloud strategies
• System design with TFX components or Kubeflow DSL (e.g., Dataflow)

5.1 summary: This section introduces the orchestration of end-to-end ML pipelines to ensure repeatability, maintainability, and scalability. You will study pipeline frameworks, pipeline hosting options, and orchestration tools such as Kubeflow Pipelines or Vertex AI Pipelines.

Practical topics include monitoring system triggers, handling hybrid deployments, and automating validations between training and serving to reduce drift. Learning these concepts helps streamline ML workflows across dynamic business environments.

5.2 Automating model retraining. Considerations include:

• Determining an appropriate retraining policy
• Continuous integration and continuous delivery (CI/CD) model deployment (e.g., Cloud Build, Jenkins)

5.2 summary: This section focuses on how to implement retraining and pipeline automation for models that evolve continuously. You will learn to define retraining policies that reflect both technical requirements and business timelines.

Alongside retraining, you will gain insights into CI/CD deployment approaches. By incorporating these practices, you can deliver reliable ML models that quickly adapt to changing data patterns.

5.3 Tracking and auditing metadata. Considerations include:

• Tracking and comparing model artifacts and versions (e.g., Vertex AI Experiments, Vertex ML Metadata)
• Hooking into model and dataset versioning
• Model and data lineage

5.3 summary: This section highlights the importance of metadata in professional ML engineering. You will study tools for comparing artifacts, managing versions, and understanding the lineage of both data and models.

By implementing metadata tracking, your ML solutions become auditable, transparent, and easier to refine. These practices provide clarity and trust within organizational deployments.

Domain 6: Monitoring AI solutions (13% of the exam)

6.1 Identifying risks to AI solutions. Considerations include:

• Building secure AI systems by protecting against unintentional exploitation of data or models (e.g., hacking)
• Aligning with Google’s Responsible AI practices (e.g., monitoring for bias)
• Assessing AI solution readiness (e.g., fairness, bias)
• Model explainability on Vertex AI (e.g., Vertex AI Prediction)

6.1 summary: This section emphasizes the need to align models with responsible AI practices while protecting them from misuse. You will assess fairness, bias, and model explainability for trustworthy outcomes.

Additionally, strategies for securing AI systems will be covered, helping ensure resilience against both technical and ethical pitfalls. This ensures solutions are reliable and aligned with organizational values.

6.2 Monitoring, testing, and troubleshooting AI solutions. Considerations include:

• Establishing continuous evaluation metrics (e.g., Vertex AI Model Monitoring, Explainable AI)
• Monitoring for training-serving skew
• Monitoring for feature attribution drift
• Monitoring model performance against baselines, simpler models, and across the time dimension
• Monitoring for common training and serving errors

6.2 summary: This section focuses on continuously evaluating model accuracy and stability. You will learn how to implement model monitoring tools that assess drift, skew, and other performance challenges affecting predictions.

The section also highlights monitoring model effectiveness using baselines and simpler models. This allows engineers to refine AI over time and ensure its predictions remain accurate, accountable, and beneficial for stakeholders.