Google Cloud Certified Professional Machine Learning Engineer Training

This course prepares you for the N/A certification exam, covering all official exam domains and their approximate weightings:

Domain 1 Architecting low-code AI solutions (13%)

• Build models in BigQuery ML or Gemini Enterprise Agent Platform AutoML (classification, regression, forecasting, clustering) based on the business problem
• Perform feature engineering/selection using BigQuery ML; generate predictions using BigQuery ML
• Train models using Agent Platform AutoML; fine-tune Gemini models using BigQuery
• Evaluate and select the appropriate model from Agent Platform Model Garden for a given task
• Build applications using industry-specific APIs (Document AI, Vision, Translate) and tune models for specific use cases (Gemini, Imagen, Veo)
• Optimize Gemini-based applications for cost, latency, and availability

Domain 2 Collaborating within and across teams to manage data and models (16%)

• Organize and explore tabular, text, and image data for efficient experimenting, training, and serving
• Choose the right preprocessing tool by scale/complexity (BigQuery SQL, Dataflow, Apache Spark, in-memory Python frameworks)
• Create/consolidate features in Agent Platform Feature Store; ensure data privacy and handle PII
• Prototype models in Agent Platform Workbench or Colab Enterprise notebooks using PyTorch, sklearn, JAX and Model Garden models
• Choose the right environment for experimentation (Experiments on Agent Platform, Agent Platform Pipelines, Kubeflow Pipelines)
• Evaluate predictive and gen AI solutions (metrics, LLM-as-a-judge) and track model artifacts/versions/lineage

Domain 3 Scaling prototypes into ML models (21%)

• Choose model type (ARIMA, DNN, LLM), product (AutoML, BigQuery ML, Pipelines), deployment strategy, and modeling technique given interpretability needs
• Organize and ingest structured/unstructured training data (Cloud Storage, BigQuery) into training pipelines
• Train models via different SDKs (Agent Platform custom training, Kubeflow on GKE, AutoML, Tabular Workflows)
• Troubleshoot ML model training failures and perform hyperparameter tuning
• Fine-tune foundational models from Agent Platform/Model Garden and judge when tuning is warranted
• Evaluate compute/accelerator options (CPU, GPU, TPU) and distributed training strategies (data/model parallelism)

Domain 4 Serving and scaling models (20%)

• Deploy models for batch and online inference (Agent Platform, Model Garden, Cloud Run, GKE)
• Package/serve models from different frameworks (PyTorch, XGBoost) using prebuilt/custom containers
• Organize and version models in Gemini Enterprise Agent Platform Model Registry
• Implement rollout strategies (A/B testing, canary deployments) and inference pre/postprocessing
• Manage and serve features via Agent Platform Feature Store; deploy to public/private endpoints
• Choose appropriate serving hardware (CPU/GPU/TPU/edge) and scale serving backend based on throughput

Domain 5 Automating and orchestrating ML pipelines (18%)

• Validate data and models within end-to-end ML pipelines
• Build and orchestrate pipelines using managed/unmanaged services or templates (Agent Platform Pipelines, Managed Service for Apache Airflow, Ray on Agent Platform)
• Ensure consistent data preprocessing between training and serving
• Determine an appropriate model retraining policy
• Deploy models in CI/CD/CT pipelines (e.g., Cloud Build)

Domain 6 Monitoring AI solutions (13%)

• Build secure AI systems against data/model exfiltration, malicious prompting, and sensitive-data leakage to LLMs (Regex, safety filters, Model Armor)
• Align with responsible AI practices, including monitoring for bias
• Provide model explainability on Agent Platform (e.g., Agent Platform Inference)
• Configure Model Monitoring on Gemini Enterprise Agent Platform for continuous evaluation metrics
• Monitor for training-serving skew, data drift, concept drift, and feature attribution drift
• Monitor, test, and evaluate gen AI solutions