Domain 1: Architecting Low-Code AI Solutions¶
Official exam weight: approximately 13%
This domain tests your ability to choose and build low-code AI solutions that fit a business problem. It is broader than MLOps: you must recognize when SQL-based ML, AutoML, a prebuilt AI API, or a foundation model is the right approach.
Official objectives¶
1.1 BigQuery ML and AutoML on Gemini Enterprise Agent Platform¶
Prepare to:
- Match classification, regression, forecasting, and clustering tasks to appropriate model types.
- Perform feature engineering and feature selection in BigQuery ML.
- Generate predictions with BigQuery ML.
- Train models using Agent Platform AutoML.
- Fine-tune Gemini models using BigQuery.
1.2 Google Cloud AI APIs and foundation models¶
Prepare to:
- Select models from Gemini Enterprise Agent Platform Model Garden.
- Choose industry APIs such as Document AI, Vision, and Translate.
- Build or tune solutions using Gemini, Imagen, Veo, and models as a service in Model Garden.
- Balance cost, latency, and availability in Gemini-based applications.
Guided labs and practice¶
| Status | Chapter | Exam objective |
|---|---|---|
| Available | Build a multiclass classification model with BigQuery ML | 1.1 |
| Planned | BigQuery ML regression, forecasting, and clustering | 1.1 |
| Planned | Feature engineering and feature selection with BigQuery ML | 1.1 |
| Planned | Train a tabular model with Agent Platform AutoML | 1.1 |
| Planned | Fine-tune a Gemini model using BigQuery | 1.1 |
| Planned | Select and invoke a foundation model from Model Garden | 1.2 |
| Planned | Build an AI API solution with Document AI or Vision | 1.2 |
| Planned | Evaluate Gemini application cost, latency, and availability | 1.2 |
Official training¶
Open the Google Skills PMLE learning path. Select activities relevant to low-code ML, BigQuery, AutoML, and foundation models. The Skills path contains 17 activities overall; it is not divided into these exact six exam-domain pages.
Knowledge checkpoint¶
Before moving on, make sure you can explain:
- When BigQuery ML is a better fit than a custom training job.
- How classification, regression, forecasting, and clustering differ.
- When to choose an industry AI API versus a foundation model.
- How cost, latency, and availability influence model selection.
📬 DevopsPilot Weekly — Learn DevOps, Cloud & Gen AI the simple way.
👉 Subscribe here