Google PROFESSIONAL MACHINE LEARNING ENGINEER Exam Prep
Professional Machine Learning Engineer (Page 21 )

1. Create a new AMI or copy the existing AMI while specifying EBS encryption and an appropriate AWS KMS key.
2. This creates an encrypted AMI backed by encrypted EBS snapshots.
3. Copy the encrypted AMI to the destination Region, using a KMS key available there.
4. Launch the application stack from the encrypted AMI.

• B: AWS KMS does not encrypt an existing unencrypted AMI in place. You must create an encrypted copy.
• C: AWS Certificate Manager manages TLS certificates, not EBS or AMI encryption.
• D: Enabling encryption by default affects newly created EBS volumes and snapshots; it does not encrypt an existing unencrypted AMI during copying.

• Multi-factor authentication
• A secure password change for users whose credentials may be compromised

• Disabling GPS: Affects location services, not email connectivity.
• Configuring 4G usage only: Does not enable international data access and could prevent connectivity if 4G is unavailable.
• Connecting using a hotspot: Could provide data, but it is not the most direct solution and depends on another network or device.

• Co to znaczy: transakcje zachowuja sie tak, jakby byly wykonywane jeden po drugim. Dzieki temu nie ma brudnych odczytów, niepowtarzalnych odczytów ani phantom reads.

• Implementacja: PostgreSQL uzywa SSI (Serializable Snapshot Isolation). W praktyce oznacza to, ze jesli wykryje konflikt miedzy transakcjami, jedna z nich zostanie zakonczona bledem “serialization failure” i trzeba ja ponowic.

• Zalózmy tabele orders(customer_id int, amount int).

• Transakcja T1:

• Transakcja T2, równoczesnie:

• T1 ponownie wykonuje to zapytanie:

• Transact-SQL (T-SQL) is used to query relational data in Azure SQL Database, not telemetry data from monitoring services.
• To analyze monitoring data from Azure SQL Database Intelligent Insights and Azure Application Insights, you should use the Kusto Query Language (KQL) in Azure Monitor Logs/Log Analytics (and in Application Insights Analytics).
• Using KQL lets you query telemetry tables such as requests, dependencies, pageViews, and customEvents (from App Insights) and any integrated monitoring data you’ve ingested into Log Analytics.

• Ensure Intelligent Insights and Application Insights data is available in a Log Analytics workspace or accessible via Application Insights Analytics.
• Go to Azure Monitor > Logs (or Application Insights > Analytics) and write KQL queries, e.g.:

• Drop violating frames
• Log or alert on the violation
• Shut down or restrict the port

• Load balancer: Distributes application traffic; it does not control switch MAC learning.
• IPS: May detect some network attacks but is not the primary control for switch MAC-table exhaustion.
• NGFW: Protects traffic crossing network boundaries, while MAC flooding occurs at the switching/LAN layer.

• Forecasting lets sales leaders predict revenue for upcoming fiscal periods, compare forecasts with targets, and review performance by salesperson or team.
• Opportunity Stages help determine the likelihood that opportunities will close. Each stage can have a probability, which contributes to forecast calculations.

• Sales Quotes: Used to present pricing or product proposals to customers; they do not provide revenue forecasting.
• Opportunity List View: Helps users filter and view opportunities, but it is not the forecasting tool.

• Source: on-premises SQL Server Database1
• Destination: Fabric Warehouse Warehouse1
• Connectivity: on-premises data gateway

• Dataflow Gen1: Primarily for Power Query transformations and isn’t the best Fabric-native choice for this copy operation.
• KQL queryset: Queries data; it does not orchestrate copying from SQL Server.
• Notebook: Could perform custom code-based movement, but it is unnecessary and less direct than a pipeline.

• Server — provides services or resources
• Endpoint — a user device connected to the network
• Ethernet — a networking technology, not a filtering device
• Firewall — filters and controls traffic

• Correct answer: B — From the code repository, modify Startup.cs (in ASP.NET Core you configure the Application Insights SDK in code, not with ApplicationInsights.config).

• Why the other options aren’t correct:

• How to implement fixed-rate sampling (ASP.NET Core):

var builder = WebApplication.CreateBuilder(args); 


    builder.Services.Configure<TelemetryConfiguration>(telemetryConfiguration => 
    { 
        var builder = telemetryConfiguration.DefaultTelemetrySink.TelemetryProcessorChainBuilder; 
        double fixedSamplingPercentage = 10; // 10% sampling 
        builder.UseSampling(fixedSamplingPercentage); 
    }); 


    builder.Services.AddApplicationInsightsTelemetry(new ApplicationInsightsServiceOptions 
    { 
        EnableAdaptiveSampling = false 
    }); 


    var app = builder.Build();

• If the account name includes the parsed search string, show the element; otherwise hide it.
• In CSS, to show an element you can set display to a visible value (commonly 'block' for block-level elements), and to hide you set display to 'none'.

• The ternary should map true to a visible display value and false to 'none'.
• Therefore the placeholder should be replaced with 'block' : 'none' (so the code becomes: accountName.includes(parsedSearchString) ? 'block' : 'none').

• There are several typos in the snippet (e.g., quesrySelectorAll, toLOwerCase, Style.display). In a working version you’d also fix those to document.querySelectorAll, account.innerHTML.toLowerCase(), and account.style.display.

• The missing piece is to configure an event source mapping between the DynamoDB stream and the Lambda function. Enabling DynamoDB Streams alone doesn’t trigger the function—you must wire the stream to Lambda so events invoke the function.
• Note that changing StreamViewType (A) only affects what data is delivered (e.g., NEW_IMAGE) and does not trigger invocations. Options C and D are not required for invocation: you don’t need SNS, and increasing the timeout won’t enable the trigger.

• A limited lifetime
• Permissions defined by an IAM role or policy
• The ability to access AWS APIs without using long-term IAM user credentials

• Amazon API Gateway — Builds and manages APIs; it does not generally issue AWS credentials for applications.
• IAM users — Usually use long-term credentials, which are less suitable for applications.
• IAM instance profiles — Provide temporary role credentials to applications running on EC2 instances, but are specifically tied to EC2. The broader requirement described is best met by AWS STS.

• Security Development Operations (SDO): a security-focused take on DevOps, integrating security into the development pipeline.
• Service Data Object (SDO): a data representation used in some middleware or enterprise apps.
• Software-Defined Overlay (SDO): a networking overlay concept.
• Secure Data Object: a generic data-security term.

• I can’t view the images, so I can’t tell which table is labeled as A in your screenshot. But here’s how to decide:

• How to identify the fact table:

• How to identify the surrogate key in the Patient table:

• Relationship:

• Basic: billing/subscription support only; no technical support requests can be opened.
• Developer, Standard, Professional Direct, and Premier: include technical support and allow opening new support requests.

• If option A corresponds to “Developer, Standard, Professional Direct, Premier,” that is the correct choice.
• If the option text is really “Premier only,” that would be incorrect per current Azure docs.

• Automation handles repetitive tasks, and orchestration coordinates responses across tools (SIEM, SOAR, EDR, etc.).
• This allows the security team to process more events with the same number of people, improving throughput and response times.
• Analysts can focus on more complex tasks like investigation and threat hunting instead of manual triage.

• It increases complexity: automation can reduce complexity when well-implemented; complexity is not the main intended benefit.
• It removes technical debt: automation/OR focuses on current incident response efficiency, not primarily on architectural debt.
• It adds additional guard rails: governance is important but not the core value of automation/orchestration within a SIEM.

• Correct answer: B — A reservation for the server based on the MAC address.

• Why this is the best choice:

• Why the other options are less suitable:

• Quick takeaway: Use a DHCP reservation to give the file server a fixed IP while still benefiting from centralized DHCP management.

• Why: A multi-modal embedding model maps both text and images into a shared vector space, enabling cross-modal similarity and retrieval (e.g., find images matching a text query or text matching an image).

• What it does: Enables efficient similarity search using a vector database or nearest-neighbor indexing.

• Why the other options aren’t suitable:

• Quick implementation note: you’d typically pair the embeddings with a vector store (e.g., ANN index) and possibly fine-tune on domain data for better alignment across modalities.

• The correct setting to configure is the Enrollment Status Page (ESP) profile.
• Why: ESP tracks the enrollment progress and logs during Autopilot provisioning. You can enable log collection within the ESP policy so that, if ESP times out, the user can collect and share the logs with support.
• Optional enhancement: In the ESP policy (or the Autopilot diagnostics settings), you can enable:

• Automatic enrollment settings: control how devices auto-enroll, not log collection.
• Windows Autopilot deployment profile: controls deployment steps but not ESP log collection.
• Device configuration profile: applies settings but doesn’t provide ESP log collection mechanics.

• Answer: D

• The strongest security comes from enforcing access at the storage layer. An IAM policy scoped to the Cognito identity prefix restricts each user to their own S3 prefix (folder), so users can only access their own files regardless of client or server code.
• This per-user isolation reduces risk of cross-user data access and is more robust than UI validation or Lambda/API validation alone.
• Implementation note: attach a policy to the Cognito identity role that allows actions on arn:aws:s3:::your-bucket/${cognito-identity.amazonaws.com:sub}/* (per-user prefix). Ensure TLS in transit and enable at-rest encryption (SSE-S3 or SSE-KMS). Use multipart upload for large files up to 300 MB.

• Strongest encryption algorithms: Choose the strongest available encryption and integrity options, typically AES with the largest key size and a secure integrity algorithm.
• Encrypt all traffic: Select tunnel mode rather than transport mode. Tunnel mode encrypts the entire original IP packet, which is appropriate for a site-to-site VPN between branch networks.
• Secret-password authentication: Configure a pre-shared key (PSK) on both VPN concentrators. The key must match at both ends.

• Residual neural network (ResNet): Primarily used for computer vision tasks such as image classification.
• Support vector machine (SVM): Commonly used for classification and regression, not natural-language query generation.
• WaveNet: Designed mainly for generating or modeling audio and speech.

• D. Provide public endpoints to store and retrieve data

• E. Manage the infrastructure layer and the operating system

• Classifying data
• Configuring access permissions, such as IAM policies
• Selecting or configuring available encryption options, depending on the requirement

• A — conformant: Incorrect because one explicit requirement is violated.
• B — consistent: “Consistent” is not sufficient where the implementation contradicts a mandatory specification. The RPC mechanism is covered by the architecture, and the implementation differs from it.
• C — consistent: Incorrect reasoning: the RPC mechanism does have a defined architectural requirement, so conformance must be assessed.
• D — non-conformant: Correct because the implementation fails to follow the specified RPC/threading requirement.

• Honeypots: Decoy systems or services designed to attract attackers. Legitimate users should not access them, so activity on a honeypot is a strong indicator of unauthorized activity. Security teams can also observe attacker behavior.

• DNS sinkhole: Redirects requests for known malicious domains to a controlled internal server or blocks them. If an infected workstation attempts to contact command-and-control infrastructure, the sinkhole records the request and helps identify the compromised device.

• Tokenization: Protects sensitive data by replacing it with tokens.
• CI/CD: Automates software development and deployment.
• Threat modeling: Identifies potential risks before or during system design.
• Data obfuscation: Makes data harder to understand, but does not detect intruders.

• Coordinate and manage the assessment team
• Assign assessment activities
• Direct interviews, examinations, and tests
• Ensure the assessment follows the CAP

• Answer key: Lead Assessor — likely wrong
• CAP-based answer: C3PAO

• The problem: Guest users can’t create posts in the Monthly Reports channel.
• The recommended fix (answer B): Modify the Guest access settings in the Teams admin center.

• Guest access settings control what guests are allowed to do across Teams (for example, posting, chat, and channel participation). If guests are restricted from posting, enabling the appropriate guest permissions will let them post in channels like Monthly Reports.
• The other options don’t directly change guest permissions in the way that affects posting:

• Box 1: Azure Lighthouse — ??? ??????? ? ?????????? ????????? ? ?????????? Entra tenants. ??? ??????????? ??? ??? ????? ????? ? VM ? ?????? ?????????/????????.
• Box 2: Azure Monitor Agent ? ?????????????? Data Collection Rules (DCR) — AMA ???????? ??????? Windows (????????, Security Logs) ? DCR ??????????, ????? ??????? ???????? ? ???? ?????????? (? ?????? Log Analytics workspace).

• Premium file shares: contoso104 only
• Blob access tiers: contoso101, contoso102, and contoso103

• Premium Azure Files shares require a storage account with the FileStorage account kind. This is a specialized account type and is not used for blob storage.
• Blob access tiers such as Hot, Cool, and—where supported—Archive apply to blob-capable accounts, such as General-purpose v2 or BlobStorage accounts shown in the exhibit.
• contoso104 is therefore excluded from the blob-tier selection because it is the premium Files account.

• Topic: Conditional Access (CA) policy and MFA for Azure management.
• Policy name: Policy1, created with the template “Require multifactor authentication for Azure management.”
• What the template does: It enforces MFA when signing in to manage Azure resources (Azure management) for the users included in the policy.
• Who is included: The question lists Admin1, Admin2, Admin3, and Admin4 as the admins in the tenant. The template’s default scope for this scenario is to apply to those admin accounts (i.e., all four admins unless you explicitly limit the scope).

• CA policy for “Azure management” targets admin sign-ins.
• If the policy includes Admin1–Admin4, all of them must use MFA on their next sign-in (choice D).

• Dynamics 365 Remote Assist — used for remote collaboration and mixed-reality scenarios.
• Dynamics 365 Finance — supports financial and operational management.
• Dynamics 365 Business Central — designed primarily for small and medium-sized businesses’ finance and operations.
• Dynamics 365 for Phones — provides mobile access to Dynamics 365 apps, including Sales, through the mobile Unified Interface.

• prompt: the input or instruction given to the model
• completion: the desired response

json 
{"prompt":"Summarize this customer complaint.","completion":"The customer reports a delayed shipment."}

• B: A .txt file containing CSV-style lines is not automatically a valid fine-tuning dataset. The required structure and format must match the selected Bedrock model.
• C: Provisioned Throughput provides dedicated inference capacity; it does not fine-tune a model.
• D: Journals and textbooks might support domain adaptation or pretraining, but they do not provide the prompt-to-answer examples typically required for supervised fine-tuning.

• IaaS: You deploy and manage virtual machines, operating systems, disks, and networking. Azure manages the physical datacenter and host infrastructure.
• PaaS: Azure manages the operating system and runtime so you can focus on deploying applications, such as with Azure App Service or Azure Functions.
• SaaS: You consume a complete, provider-managed application.

• Answer: A) Clear the Windows event logs

• Windows event logs (System, Security, Application) are key forensic artifacts. If an attacker clears these logs, there’s no built-in local record of actions taken, making it harder for investigators to trace what happened.

• Modify the system time: may misalign timestamps but does not erase past events and can be easily detected.
• Alter log permissions: could prevent new log entries, but it doesn’t remove existing ones and can raise alarms due to permission changes.
• Reduce log retention settings: affects only future logs; it doesn’t wipe already recorded activity.

• In real environments, defenders mitigate this risk with tamper-evident, centralized logging (remote SIEM, write-once storage, log integrity checks) to preserve evidence even if a host is compromised.

• Correct answer: C — The model is deployed at UK border checkpoints.

• The EU AI Act applies to AI systems used within the EU or by entities established in the EU. So:

• A Storage Gateway File Gateway provides an on-premises NFS/SMB file share that's backed by objects in S3. This is exactly what you need to migrate an NFS workload: applications continue using NFS, while the actual data moves to and is stored in S3.
• It can cache frequently accessed data locally for performance, while uploads go to an S3 bucket.
• Why the other options aren’t suitable:

• Correct answer: A — a clear business case has been established.

• The audit focus for IT procurement is ensuring there is a solid economic justification for the investment. A clear business case documents the expected benefits, costs, risks, and alignment with strategic objectives, which governs whether the investment is worth pursuing.
• The other options are important controls or project aspects, but they are secondary to ensuring the investment is justified financially and strategically:

• The Act requires:

• It does not, as a public obligation, mandate publishing a detailed report on the training data used.
• Also, the Act does not specify a fixed “six months” retention for system logs in a way that would clearly make A required; logs for traceability are implied, but the six-month period is not a defined, universal requirement.

• Manage multiple organizations from one hierarchy.
• Enterprise-wide policies, including repository visibility, forking, and GitHub Copilot policies.
• Centralized identity and access management, such as SAML SSO, SCIM provisioning, and—where applicable—Enterprise Managed Users.
• Centralized billing and licensing, including organization and Copilot seat management.
• Enterprise audit log, which aggregates activity across organizations.
• Security and compliance oversight, such as organization security visibility and enterprise-wide controls.
• Enterprise support and account management features, depending on the plan.

• Organization: manages its own repositories, teams, members, and local policies.
• Enterprise: governs multiple organizations and can enforce policies across them.

• Goal: serving a static website globally with minimal latency.
• Why C works: CloudFront is a Content Delivery Network (CDN) that caches content at many edge locations around the world. When users request the site, CloudFront serves cached copies from the nearest edge location, greatly reducing latency. It also reduces load on your EC2 origin by handling many requests at the edge.
• Why the other options aren’t as good:

Flat organisation

• Few levels of management
• A wide span of control—managers supervise many employees
• Shorter communication channels
• More employee autonomy and quicker decisions
• Fewer formal promotion levels

Tall organisation

• Many levels of management
• A narrow span of control—managers supervise relatively few employees
• A clear hierarchy and reporting structure
• More formal supervision and communication channels
• More obvious promotion and career stages

• Correct answer: A) Hashing

• Why: Hashing produces a fixed-size digest of the software. If the final build is altered in any way, the hash will change, enabling you to detect tampering by comparing the stored digest with a freshly computed one.

• Why not the others:

• Answer: D. Select an IdP

• Why: To reduce the number of credentials across multiple SaaS apps, implement a single sign-on approach via an Identity Provider. The first step is choosing the IdP; once selected, you can enable SAML (A) or OAuth as supported by apps. Password vaulting (C) is not a true SSO solution, and B is not the foundational step.

• Review the risk register.
• Use the previously defined risk response plan, such as assigning a backup resource, reprioritizing work, or using another qualified engineer.
• Implement the response promptly because the task is on the critical path.

• B: Revising the schedule without first applying the planned response may be premature.
• C: Excluding a required task could compromise the project’s objectives and requirements.
• D: Updating the risk and lessons learned records is appropriate, but it should happen after managing the immediate issue, not as the next primary action.

• Correct options: C and D.

• The AWS Cloud Adoption Framework (CAF) security perspective focuses on capabilities to protect information and manage security risk.
• Observability (A) is primarily an operations/monitoring capability, not a core security capability under CAF.
• Incident and problem management (B) is IT service management/operations, not specifically a CAF security capability.
• Incident response (C) is a security-focused capability: detecting, containing, and recovering from security incidents.
• Infrastructure protection (D) involves applying security controls to protect the underlying infrastructure, which fits the security perspective.

• runParallel is defined as async, so it returns a Promise.
• To use the result, you attach a then handler (or use await inside another async function).

• A: “Async runParallel() .then(data);” is invalid. “async” is a keyword for function declarations, not something you put in front of a call, and you’d need a function as the then-argument.
• B: “runParallel().done(function(data){return data;});” .done is not standard in native Promises (it exists in some libraries). Without a compatible library, this won’t work.
• C: “runParallel().then(data);” is invalid because then expects a function as its argument; data is not a function.
• D: “runParallel().then(function(data){ return data; });” is valid (or equivalently, runParallel().then(data => data);).

• The robust, standard option is D.
• B could work only with a library that provides .done; otherwise it’s not reliable.

• The requirement is a fully managed file server that natively supports Microsoft workloads and file systems and uses the SMB protocol. FSx for Windows File Server is built specifically for Windows workloads, provides native SMB file shares, and supports Windows permissions/Active Directory integration.
• Why the others don’t fit: