Google Cloud: Vertex AI, Cloud Run, and Cloud Functions
Words made simple
A few useful meanings before you read the explanation.
- AI
- Software techniques for tasks such as recognizing patterns, understanding language or generating content.
- API
- A way for one piece of software to ask another for data or an action. A claims API might return the current status of a claim.
- Inference
- Using a model that has already been trained to produce an answer or prediction.
More words from this lesson
- Vector
- An ordered list of numbers. An embedding is a vector used to represent content for tasks such as similarity search.
- Agent
- An application that uses a model to choose its next action within defined limits. It can use tools and continue through several steps.
- Authorization
- Checking what that identity is allowed to read or do. Being logged in does not grant access to every record.
- Evaluation
- Checking results against examples and clear criteria to find whether the system works well and where it fails.
- Observability
- Being able to understand system behavior through records such as logs, measurements and traces.
Overview
- Vertex AI is the familiar Google Cloud AI platform name used in many existing projects; its model documentation currently routes into evolving Google enterprise agent-platform documentation.
- Cloud Run executes containerized applications and workloads.
- Cloud Functions terminology appears within Cloud Run functions documentation for function-style deployment.
Why this matters
Google Cloud teams can connect supported models to existing identity, storage, and container execution.
An everyday example
Use the company's existing office and access system while hiring an approved language specialist.
A technical example
A Cloud Run Node API calls an approved model; a separate parsing service processes document jobs.
How it works
- Establish service identity
- Deploy compute
- Configure approved model endpoints
- Process/retrieve evidence
- Validate output
- Emit traces/metrics.
Architecture and flow
Authenticated app → Cloud Run/functions → model platform + search/storage services.
The insurance assistant example
A Google-centered insurer deploys Node and Python containers with controlled service-to-service access.
When to use it
Existing Google Cloud operations and evaluated model/deployment fit.
When to avoid it
Do not confuse the developer Gemini API account/configuration with every Google Cloud model-service feature or governance setting.
Interview answer
“I map Google's model platform and container/function services to the same inference, compute, identity, and data boundaries used on other clouds.”
Benefits, trade-offs, and production considerations
- Verify current product names, endpoints, regions, quotas, service identities, and networking before implementation.
- These change faster than the architecture.
- Model platform docs, Cloud Run, Cloud Run functions.
| Requirement | Azure approach | AWS approach | Google Cloud approach |
|---|---|---|---|
| Hosted models | Azure OpenAI / Foundry model services | Bedrock | Vertex/Google model-platform services |
| Container services | Container Apps | ECS | Cloud Run |
| Event functions | Functions | Lambda | Cloud Run functions |
| Knowledge search | AI Search | OpenSearch or chosen vector store | Chosen managed/search/vector service |
| Identity/secrets | Entra/managed identity + Key Vault | IAM + approved secret management | Service identities/IAM + approved secret management |
| Primary decision | Existing enterprise fit, approved regions, model features, total operating cost | Same criteria | Same criteria |
- A cloud product diagram is not a finished system design.
- You must still specify document lifecycle, business authorization, retries, budgets, evaluation, observability, and disaster recovery.
Can you explain it without the notes?
Google Cloud: Vertex AI, Cloud Run, and Cloud Functions
Use your own words. Explain an example and one limitation before checking the explanation.
Check your explanation
Vertex AI is the familiar Google Cloud AI platform name used in many existing projects; its model documentation currently routes into evolving Google enterprise agent-platform documentation. Cloud Run executes containerized applications and workloads. Cloud…
Turn this concept into a clear answer.
Practice this conceptAdapted from 13_AI.MD · “Google Cloud: Vertex AI, Cloud Run, and Cloud Functions”. Examples are study fragments; framework behavior can depend on your version.