Practice Exams:

Amazon

Amazon AWS SAP-C02: Enterprise AWS Cost Optimization Starts With Ownership

Cloud cost problems are rarely caused by one expensive service. They are usually caused by a system in which nobody can clearly explain who owns a workload, what business outcome it supports, why its consumption changed, or who is authorized to trade performance and resilience for lower spend. At enterprise scale, cost optimization begins as an ownership problem before it becomes a pricing problem. Home/Amazon/Amazon AWS SAP-C02: Enterprise AWS Cost Optimization Starts With OwnershipAmazonPublished October 3, 2026On this pageCost allocation is useful only when it maps to real responsibilityA cost…

Read More

Amazon AWS SAP-C02: Designing for Failure Across Accounts and Regions

Reliable systems are not created by assuming infrastructure will stay healthy. They are created by deciding which failures are acceptable, which failures must be isolated, how quickly service must recover, and which dependencies are allowed to fail together. In AWS, those decisions span components, Availability Zones, Regions, accounts, identity systems, deployment pipelines, and the teams that operate them. Home/Amazon/Amazon AWS SAP-C02: Designing for Failure Across Accounts and RegionsAmazonPublished October 3, 2026On this pageStart with failure domains, not a list of AWS servicesMulti-AZ is usually the first resilience boundary inside a…

Read More

Amazon AWS AIF-C01: RAG vs. Fine-Tuning vs. Prompting

Retrieval-augmented generation, fine-tuning, and prompt engineering are often presented as competing ways to “improve” a generative AI system. They are better understood as tools for different failure modes. Prompting changes the instructions and context given at inference time. RAG supplies external information that the model can use for a response. Fine-tuning changes model behavior by training it on examples. Choosing the wrong technique can add cost and complexity without solving the real problem. Home/Amazon/Amazon AWS AIF-C01: RAG vs. Fine-Tuning vs. PromptingAmazonPublished October 3, 2026On this pagePrompt engineering changes the request,…

Read More

Amazon AWS AIF-C01: Foundation Models for Business Use Cases

Choosing a foundation model is not the same as choosing the model with the most impressive benchmark or the largest parameter count. A business application has a job to perform, users to serve, data it may or may not be allowed to access, latency and cost constraints, and a tolerance for error. Model selection should begin with those requirements and work backward to the capabilities that matter. Home/Amazon/Amazon AWS AIF-C01: Foundation Models for Business Use CasesAmazonPublished October 3, 2026On this pageStart with the business task, not the model catalogFoundation model,…

Read More

Amazon AWS AIF-C01: Responsible AI Is a Product Requirement

Responsible AI should not appear at the end of a project as a policy document that nobody used to shape the product. Fairness, explainability, privacy, security, safety, controllability, robustness, and governance affect requirements, architecture, data choices, evaluation, user experience, monitoring, and release decisions. If those concerns are postponed until deployment, the expensive parts of the system may already be difficult to change. Home/Amazon/Amazon AWS AIF-C01: Responsible AI Is a Product RequirementAmazonPublished October 3, 2026On this pageStart by defining who can be harmed and howFairness needs use-case-specific evidenceExplainability should match the…

Read More

Amazon AWS AIF-C01: Bedrock Guardrails: Where Safety Controls Fit

Amazon Bedrock Guardrails are most useful when they are treated as one layer in an application’s safety architecture rather than as a switch that makes generative AI safe. A guardrail can evaluate user input and model output against configured policies, block or mask content, and support consistent controls across supported models and workflows. It cannot decide the business risk tolerance, fix poor authorization, guarantee factual accuracy, or replace application testing. Home/Amazon/Amazon AWS AIF-C01: Bedrock Guardrails: Where Safety Controls FitAmazonPublished October 3, 2026On this pageGuardrails evaluate content around model inferenceContent filters…

Read More

Amazon AWS AIF-C01: GenAI Security Starts With Data, Identity, and Access

Generative AI introduces new attack surfaces, but many of the most damaging failures still begin with familiar security problems: data was accessible to the wrong identity, credentials were over-privileged, secrets were embedded in prompts, logs exposed sensitive content, or an application allowed a model to call a tool without independent authorization. Treating GenAI security as a completely new discipline can distract teams from the controls they already know how to apply. Home/Amazon/Amazon AWS AIF-C01: GenAI Security Starts With Data, Identity, and AccessAmazonPublished October 3, 2026On this pageClassify the data before…

Read More

Amazon AWS AIF-C01: Evaluating AI Outputs Beyond a Single Accuracy Score

Generative AI evaluation becomes misleading when every problem is reduced to one “accuracy” number. A model can be factually correct but irrelevant, helpful but unsafe, well grounded but too slow, concise but incomplete, or excellent on common prompts while failing badly on an important minority of cases. Product quality is multidimensional, so evaluation has to reflect the real task. Home/Amazon/Amazon AWS AIF-C01: Evaluating AI Outputs Beyond a Single Accuracy ScoreAmazonPublished October 3, 2026On this pageStart with a task-specific definition of successReference-based metrics work only when a meaningful reference existsHuman evaluation…

Read More

Amazon AWS AIF-C01: Choose AWS AI Services by Use Case, Not Hype

AWS offers a wide AI and machine learning portfolio, which makes service selection easy to overcomplicate. The simplest decision rule is to start with the business task and choose the highest-level managed service that meets the requirement. Use a specialized AI API when the job is well defined, Amazon Bedrock when the application needs foundation models and generative AI capabilities, and Amazon SageMaker AI when the organization needs deeper control over building, training, customizing, or operating machine learning models. Home/Amazon/Amazon AWS AIF-C01: Choose AWS AI Services by Use Case, Not…

Read More

Amazon AWS AIP-C01: Agents Need Boundaries More Than They Need More Tools

Home/Agents need boundaries more than they need more toolsAn AI agent becomes interesting when it can do something, not merely say something. The moment a model can call a search service, open a ticket, update a record, invoke a function, or trigger a workflow, the application crosses an architectural boundary. The problem is no longer limited to whether the model generates a good answer. It now includes whether the system can take the wrong action, at the wrong time, with the wrong authority.On this pageEvery new tool expands the agent’s…

Read More

Amazon AWS AIP-C01: Evaluating GenAI Without Grading Yourself

Home/Evaluating generative ai without grading your own homeworkGenerative AI systems are unusually easy to overestimate. A team builds the prompts, chooses the model, selects the demonstrations, and then reviews examples that came from the same assumptions. The output sounds fluent, the demo answers familiar questions, and everyone concludes that quality is high. That is not evaluation. It is a feedback loop in which the system is being graded by the people who already know how it was intended to behave.On this pageStart with the decision the evaluation must supportBuild representative…

Read More

Amazon AWS AIP-C01: Prompt Management as an Engineering Problem

Home/Prompt management becomes an engineering problem at scaleA prompt can begin life as a few lines in a notebook and end up controlling a business process used by thousands of people. That transition is where informal prompt writing stops being enough. Once prompts influence production behavior, they become configuration artifacts with dependencies, versions, owners, tests, deployment history, and rollback requirements.On this pageProduction prompts extend beyond instructionsVersioning creates a history operators can reason aboutPrompt variables need schemas and ownershipChange control should connect prompts to evaluationSeparate prompt development from prompt deploymentSeparate model…

Read More

Amazon AWS AIP-C01: Guardrails, Moderation, and Model Safety Limits

Home/Guardrails moderation and the limits of model safety controlsSafety controls are easiest to misunderstand when they work most of the time. A content filter blocks an obviously harmful request, a sensitive-information policy masks a phone number, and the team begins to treat “guardrails enabled” as a complete safety architecture. The dangerous gap is everything those controls were never designed to decide.On this pageContent moderation solves a classification problemApply sensitive-information filters correctlyPrompt-attack detection reduces risk but cannot define authorityDenied topics are policy controls, not reasoning enginesGuardrails do not replace deterministic business…

Read More

Amazon AWS AIP-C01: Designing GenAI for Cost Before the Bill Arrives

Home/Designing genai applications for cost before the bill arrivesGenerative AI cost problems rarely begin with an unexpectedly expensive model. They begin with an architecture that never defined what one useful outcome should cost. A prototype sends a long prompt, retrieves too much context, calls the model several times, retries on uncertainty, and then adds an agent loop. Each choice seems small in isolation. At production volume, the choices multiply.On this pageDefine a cost unit that matches the productToken economics start with prompt architectureChoose the smallest adequate modelRetrieval can multiply token…

Read More

Amazon AWS AIP-C01: AI Observability Beyond Latency

Home/Observability for ai systems what to measure beyond latencyA generative AI endpoint can be fast, return HTTP 200, and still be failing. It may retrieve the wrong evidence, use three times the expected tokens, choose an unnecessary tool, violate a policy, cite stale sources, or generate a fluent answer that users immediately correct. Traditional service metrics capture availability and performance; they do not capture whether an AI system is behaving well.On this pageStart with the conventional service health signalsTokens measure cost and behaviorQuality needs observable proxies and evaluation loopsRAG systems…

Read More