Foundations to production RAG and multi-agent systems. Each track shows difficulty, volume, prerequisites, related services, and its interview / certification / production relevance.
Cloud computing models, service models, and how Google Cloud is organized: projects, billing, regions, and zones.
Organizations, folders, projects, org policies, labels, tags, and multi-team governance patterns.
Principals, roles, policies, conditional & deny policies, Workload Identity Federation, and least privilege.
VPC, subnets, firewall policies, Cloud NAT, load balancing, Private Service Connect, and zero-trust networking.
VMs, machine families, managed instance groups, autoscaling, autohealing, Shielded/Spot VMs, and IAP access.
Services, Jobs, concurrency, cold starts, VPC egress, service-to-service auth, and deploying FastAPI to production.
Autopilot vs Standard, workloads, autoscaling, Workload Identity, private clusters, and GPU inference workloads.
Cloud Storage buckets, storage classes, signed URLs, lifecycle, plus Persistent Disk, Hyperdisk, and Filestore.
Cloud SQL, AlloyDB (+ AI/vector), Firestore, Bigtable, Spanner, and Memorystore — with a selection framework.
REST design, FastAPI on 10 deployment paths, auth, idempotency, tracing, and private internal services.
Pub/Sub, Eventarc, Cloud Tasks, and Cloud Scheduler — push vs pull, DLQs, ordering, and idempotent consumers.
Workflows, Cloud Composer (Airflow), and a decision tool across Batch, Cloud Run Jobs, and Vertex AI Pipelines.
Async job APIs, state tables, polling vs webhooks, checkpointing, idempotency, and DLQ handling for heavy work.
Batch & streaming, data lakes/warehouses, Dataflow (Beam), Dataproc, Composer, Datastream, and Dataform.
Partitioning, clustering, query optimization & cost, BigQuery ML, security controls, and Vertex AI integration.
Structured logging, metrics, tracing, SLOs/SLIs, error budgets, and an interactive troubleshooting simulator.
Terraform with the Google provider: modules, state, drift, environments, plus landing-zone & project-factory patterns.
Cloud Build, Artifact Registry, Cloud Deploy, GitHub Actions + WIF, canary/blue-green, and supply-chain security.
Least privilege, WIF, KMS/CMEK, Cloud Armor, VPC-SC, Binary Authorization, and zero-trust multi-tenant design.
Multi-zone/region architecture, failover, RPO/RTO, retries with backoff & jitter, circuit breakers, and chaos testing.
Requirement analysis, service selection, capacity estimation, integration patterns, and architecture communication.
Budgets, billing exports, rightsizing, CUDs, Spot, unit economics, and configurable cost calculators.
ML lifecycle, custom training, tuning, endpoints, Model Registry, pipelines, monitoring, and BigQuery ML.
Model Garden, Gemini APIs, prompting, structured output, function calling, embeddings, safety, and eval.
Ingestion, chunking, embeddings, hybrid retrieval, reranking, grounding, eval, and 18 architecture patterns.
Tool calling, agent orchestration, Agent Builder / Engine / ADK, evaluation, guardrails, and agentic RAG.
Provider abstraction, model routing, fallback, semantic & prompt caching, quotas, token budgets, and guardrails.
CI/CD for ML, pipelines, model & prompt versioning, eval gates, monitoring, drift, and online evaluation.
Design questions, service-selection drills, trade-offs, failure scenarios, and strong-answer frameworks by level.
Mapped study plans for Associate Cloud Engineer, Professional Cloud Architect, DevOps, Data, and ML Engineer.
End-to-end buildable projects with repo structure, Dockerfiles, Terraform, verification, and cleanup steps.