🧠Skill assessment · 180 questions

Building AI on the Cloud skill assessment

Bedrock, Vertex, Foundry — RAG and agents in production.

Take the assessment →🧑‍🏫 Learn hands-onProfessional→

What it covers

Managed LLM APIs

Understand how hosted LLM APIs actually work — model selection, token limits, rate limits — before you touch any specific cloud. Every Bedrock, Vertex AI, or Foundry integration runs into the same quota, cost, and context-window tradeoffs.

Model selection: latency, quality, and cost tradeoffsToken accounting and context-window budgetingRate limits and quota planning (RPM/TPM)Provider abstraction and multi-vendor fallbackStreaming, function calling, and structured output

AWS Bedrock

Learn how Amazon Bedrock gives you managed access to foundation models plus the RAG and agent tooling to ship production genAI on AWS. Bedrock is AWS's default answer for teams building LLM apps without hosting their own inference stack.

Model access to Claude, Llama, Nova, and other Bedrock providersOn-demand, Provisioned Throughput, and cross-Region inferenceKnowledge Bases for Bedrock and managed RAG ingestionGuardrails for Bedrock: filters, PII redaction, denied topicsAgents for Bedrock and AgentCore for multi-step tool use

GCP Vertex AI

Understand Vertex AI's model catalog, search, and agent tooling so you can ship production genAI on GCP. Model Garden, Vertex AI Search, and RAG Engine are the core APIs you'll actually call.

Model Garden: Gemini, Claude, Llama, and open modelsVertex AI Search for turnkey grounded retrievalRAG Engine for custom retrieval pipelinesGrounding with Google Search and citation checkingAgent Builder and Agent Garden for deploying agents

Azure AI Foundry

Understand Azure AI Foundry's model catalog and agent service so you can ship production genAI on Azure. Model catalog, Agent Service, and Content Safety are what job postings and certs test.

Foundry model catalog: OpenAI, Claude, Meta, Mistral, and moreFoundry Agent Service for building and hosting agentsAzure AI Content Safety and prompt-injection filteringAzure AI Search integration for retrieval-grounded agentsFoundry Control Plane: token limits, tracing, evaluation

RAG on the Cloud

Learn how retrieval scales on managed cloud infrastructure — vector stores, chunking, and embedding pipelines — so your RAG app survives past a notebook demo. Production RAG is an infrastructure problem as much as a prompting one, and it's usually where interviews go deep.

Managed vector stores: OpenSearch, Vertex AI Vector Search, Azure AI SearchChunking strategy and chunk-size tuning for retrieval qualityEmbedding model selection and batch embedding pipelinesHybrid search: dense + sparse retrieval with RRFReranking and citation grounding for accurate answers

Production & Governance

Learn the cost controls, monitoring, and compliance guardrails that separate a genAI prototype from something you can run in production. This is what turns 'it works in the demo' into 'it survives a security review.'

Cost controls: token budgets, prompt caching, model routingObservability: tracing, logging, and evaluation for LLM appsData residency, encryption, and compliance for model endpointsAccess controls and network isolation for model endpointsGuardrails, content filtering, and red-teaming before launch

Ready to benchmark your Building AI on the Cloud skills?

180 questions · about 5 minutes · see your level and percentile instantly.

Take the assessment →

Explore other assessments

🤖AI Engineering 🧑‍💻AI for Coding ☸️Kubernetes ☁️AWS Cloud 🎨Frontend Engineering ⚙️Backend Engineering 🗄️Databases 📦Containerization 🔐Security & DevSecOps 🏗️Terraform & IaC 📊Observability & Monitoring 🐍Python 🚀CI/CD & Delivery Pipelines 🔁GitOps & Delivery 🧩System Design 🐧Linux & the Command Line 🔷Microsoft Azure 🌥️Google Cloud 📜Shell Scripting