Google Cloud Cloud Architect

Choosing where code runs shapes everything after it. You compare Compute Engine, GKE and Cloud Run, plan scaling and use of Spot VMs, and configure compute resources to match workload patterns.

• Compute Engine families, managed instance groups and Spot VMs
• GKE cluster design and modes
• Cloud Run and serverless options
• Choosing a platform from workload traits
• Autoscaling and capacity planning
• Provisioning compute with templates

You build A compute design for four workloads that names a platform and scaling plan for each.

Data decisions are the hardest to reverse. You match Cloud Storage, block and file storage, relational, NoSQL and analytic services to access patterns, consistency needs and growth.

• Cloud Storage classes and lifecycle rules
• Block and file storage options
• Cloud SQL, AlloyDB and Spanner for relational data
• Bigtable and Firestore for NoSQL workloads
• BigQuery for analytics
• Configuring storage for performance and cost

You build A data tier design that matches a store to each access pattern in a case study.

Networks tie every design together and are costly to redo. You plan VPC structure and addressing, connect on-premises and other clouds, and configure load balancing and DNS for the traffic the design expects.

• VPC design, subnets and Shared VPC
• Cloud Interconnect and VPN for hybrid links
• Multicloud connectivity patterns
• Load balancer selection and configuration
• DNS and name resolution design
• Private access to Google services

You build A network design that connects a data center, two projects and a second cloud.

Every architecture balances reliability against spend. You choose regional or multi-regional layouts, set targets, apply cost levers like committed use and Spot capacity, and plan for how the solution will improve later.

• Zonal, regional and multi-regional design
• Service level objectives as design targets
• Committed use discounts and Spot capacity
• Performance bottlenecks and caching
• Cost optimization levers
• Planning future improvements

You build A cost and availability comparison of three designs for one workload.

AI workloads are now part of ordinary architecture briefs. You place Gemini models and the AI platform in a design, use prebuilt APIs where they fit, and choose accelerators and infrastructure for training and serving.

• Gemini models and the Vertex AI platform in a design
• Prebuilt AI APIs versus custom models
• AI Hypercomputer and accelerator choices
• Data flow for ML workflows
• Serving patterns and capacity for AI
• Cost and quota planning for AI workloads

You build An architecture that adds a generative AI feature to an existing application, with capacity and cost noted.

A good hierarchy makes policy simple. You structure organizations, folders and projects, design IAM and service account use, and apply organization policies so teams move quickly inside safe limits.

• Organization, folders and projects
• IAM roles and least privilege
• Service accounts and workload identity
• Organization policy constraints
• Separating duties across environments
• Federating workforce identities

You build A resource hierarchy and IAM model for three business units.

Protecting data means controlling where it can go as well as who can read it. You design encryption and key use, perimeters with VPC Service Controls, and safeguards for AI such as screening prompts with Model Armor.

• Encryption at rest and in transit
• Cloud KMS and customer-managed keys
• VPC Service Controls perimeters
• Sensitive data discovery and protection
• Secret management
• Model Armor and other controls for AI

You build A data protection design with key ownership, a perimeter and AI controls.

Regulations turn into architecture constraints. You map legal and industry requirements to controls, apply data sovereignty and regional rules, and use the shared responsibility model to show who owns what.

• Mapping regulations to technical controls
• Data sovereignty and regional placement
• Shared responsibility on Google Cloud
• Assured Workloads and compliance tooling
• Audit evidence and logging requirements
• Compliance for AI uses

You build A compliance control map for a regulated workload.

Moving to the cloud changes teams and budgets as much as servers. You plan migrations in phases, pick tools for data and workloads, and address stakeholders, skills and the shift from capital to operating costs.

• Assessing the estate and choosing migration approaches
• Planning phases, dependencies and cutover
• Tools for moving data, databases and virtual machines
• Stakeholders, communication and change management
• Skills and operating model changes
• CapEx versus OpEx and what finance needs to see

You build A phased migration plan with a stakeholder and skills plan.

Analyzing technical processes means looking at how change ships and how recovery works. You review CI/CD, testing practice and disaster recovery designs, and judge them against stated recovery targets.

• CI/CD design and release strategies
• Testing strategy across environments
• Disaster recovery patterns and targets
• Backup and restore design
• Business continuity planning
• Reviewing a process for gaps

You build A review of a delivery and recovery process with the changes listed in the order to apply them.

Architects guide those who build. You advise on infrastructure as code, API management and deployment approaches, use gcloud and SDKs for repeatable work, and plan load tests that prove a design holds.

• Terraform for repeatable environments
• gcloud and client libraries in automation
• API management with Apigee
• Deployment and testing guidance for developers
• Load testing and capacity checks
• Guiding teams through design reviews

You build A Terraform baseline and an API management plan for a shared platform.

A design is only as good as its running behavior. You apply observability, alerting, release management and quality control, and test reliability with fault injection and security testing.

• Monitoring and logging design
• Alerting that reflects user impact
• Release management and change control
• Chaos engineering and reliability tests
• Penetration testing and security validation
• Incident response and learning from failures

You build An operational readiness review with monitoring, alerting and a reliability test plan.

Work a full scenario from brief to defense. You analyze the company, design compute, data, network, security and operations, estimate cost and risk, and answer objections from stakeholders who disagree.

• Analyzing a new business scenario
• Designing the solution end to end
• Covering security, compliance and AI
• Costing, phasing and migrating
• Presenting and defending the architecture

You build A complete architecture proposal with diagrams, costs, risks and a migration plan.