CFA CFA I Exam Prep
CFA Level I Chartered Financial Analyst (Page 278 )

• Azure Cosmos DB is a globally distributed NoSQL database service. It stores and retrieves application data, but it does not provide the tools needed to build, test, and deploy predictive analytics models.
• Azure Machine Learning—including the historical Azure Machine Learning Studio—is designed for the machine-learning lifecycle:

• Compensating control: The server is end-of-life, so it may no longer receive security patches or support. Placing it on a restricted VLAN provides an alternative layer of protection by reducing who can access it and limiting its exposure.
• Technical control: VLAN segmentation and access control lists (ACLs) are technology-based mechanisms that enforce network access restrictions automatically.

• Managerial: Policies, procedures, and governance—not network configuration.
• Physical: Locks, guards, or barriers.
• Corrective: Restores systems after an incident.
• Detective: Identifies suspicious activity rather than restricting access.
• Deterrent: Discourages behavior but does not enforce access rules.

• Input feature distributions differ from training data ? Analyze dataset drift metrics

• Model accuracy drops without code changes ? Analyze dataset drift metrics

• Endpoint latency increases under load ? Investigate scaling and infrastructure metrics

• R-squared (R²): mede quanto da variação dos valores reais é explicada pelo modelo. É mais usado em regressão, não em classificação.

• Accuracy (acurácia): mede a proporção de previsões corretas em relação ao total. É adequada para a classificação de imagens da questão.

• RMSE (Root Mean Squared Error): mede o tamanho médio dos erros entre valores previstos e reais, penalizando erros maiores. É usado principalmente em regressão.

• Learning rate (taxa de aprendizado): controla o tamanho dos ajustes feitos pelo modelo durante o treinamento. Não é uma métrica de avaliação; é um hiperparâmetro.

• Drops packets when traffic exceeds a certain bitrate: This describes traffic policing, not shaping.
• Fragments large packets and queues them: Shaping does not fragment packets as its primary congestion-control action.
• Drops packets randomly from lower-priority queues: Packet dropping may occur if a shaping queue becomes full, but it is not the defining behavior.

• Shaping: buffers and delays excess traffic.
• Policing: drops or marks traffic that exceeds the configured rate.

1. Which host originated the infection

1. Which remaining hosts are infected

• The connecting host keeps the private key securely.
• The destination host stores the corresponding public key in its authorized keys.
• Authentication can occur without a person entering credentials each time.
• SSH also provides encrypted communication and host authentication.

• Biometric: Requires a human user and is difficult to automate.
• Smart card: Often requires user interaction or card access.
• Hard-coded password: Unsafe because passwords stored in scripts or applications can be exposed, reused, or difficult to manage securely.

• Separate accounts/domains create a strong administrative boundary between non-production and production.
• Separate projects provide distinct IAM boundaries, billing, quotas, service accounts, and resources.
• Developers can receive permissions in the development project without automatically gaining access to production.
• Production administrators can be restricted from development resources, supporting separation of duties and least privilege.

• A: Creates excessive project fragmentation without being required by the stated requirement.
• C: Keeps development and production under one account and does not clearly separate administrative duties.
• D: Uses one account and only two projects; it provides some project isolation but weaker organizational separation.

sql 
SELECT COUNT(*) FROM table;

• A: It does not normally need to scan all data files for this operation.
• B: Results are not simply reused from a query cache every time. Caching is separate from Delta’s metadata optimization.
• D: The relevant metadata is maintained through the Delta transaction log, not directly treated as a Parquet-file metadata operation in this context.

Why

cmd 
PowerShell.exe -ExecutionPolicy Bypass -File C:\Scripts\ConfigureVM.ps1

Why the other options are less suitable

• GPO logon script: Runs only when a user logs on and may not run during initial VM provisioning.
• GPO startup script: Can work, but requires the VM to be joined to the domain and the policy to be applied. It is also normally evaluated at each startup.
• Place scripts in a VHD: Merely placing scripts on a disk does not cause them to execute.

• A — Add a Path for Stage: A Path lets users update the Opportunity’s Stage directly from the record page, without opening the full edit form. It can also display guidance and key fields.
• B — Use a Kanban list view: Users can move Opportunities between stages by dragging records between columns, which is ideal when Stage is the only field they normally change.

• Reviewing and validating generated code
• Checking security, licensing, and quality concerns
• Testing the implementation
• Making the final design and coding decisions

• A: Copilot is not a complete replacement for developers.
• C: Human input, review, and oversight are still required.
• D: Copilot Chat supports many tasks beyond debugging, including explanation, generation, refactoring, and testing assistance.

• 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.

1. Ingestion Lambda receives requests from API Gateway and places messages in an SQS queue.
2. Processing Lambda is triggered by the queue and loads records into Aurora PostgreSQL.

• Decoupling: API request handling is separated from database loading.
• Buffering: SQS absorbs bursts when incoming traffic exceeds the database’s processing capacity.
• Controlled concurrency: Configure the SQS-triggered Lambda’s concurrency and batch size to avoid overwhelming Aurora or exhausting database connections.
• Automatic retries: Messages that fail processing can return to the queue and be retried. A dead-letter queue can capture persistent failures.
• Lower operational effort: The solution remains serverless and avoids managing EC2 servers or Kubernetes clusters.

• A: EC2 and Tomcat add server-management overhead.
• B: DynamoDB/DAX would require changing the database model and does not solve the Aurora ingestion design.
• C: SNS is useful for notifications and fanout, but SQS is better for durable buffering and backpressure in this producer-consumer workflow.

• Conserve public IPv4 addresses: IPv4 addresses are limited, so many internal devices share a smaller number of public addresses through NAT.
• Reduce direct exposure: Private addresses are not directly routable from the public Internet, which helps reduce unsolicited access to internal devices.

• 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.

json 
"Resource": "arn:aws:s3:::bucket-name/*"

• A: Revoking or changing an instance profile does not reliably invalidate temporary credentials that have already been issued and may still be valid.
• B: Changing the EC2 security group does not necessarily stop the attacker from using the stolen credentials from elsewhere.
• C: Amazon Macie helps discover sensitive data; Object Lock prevents modification/deletion, not reads or downloads.

• Users logging in more frequently than average
• Logins occurring during non-business hours
• Patterns that may indicate compromised credentials or suspicious access

• Login History: Shows individual login attempts and basic details, but does not provide behavioral analysis or averages.
• Login Inspector: Used to troubleshoot or inspect a specific login attempt, rather than analyze organization-wide patterns.
• Login Report: May provide reportable login information, but it is not the specialized tool for forensic behavioral analysis.

• Users retain their existing usernames and passwords.
• Existing accounts can be used to access Azure-based resources.
• Administrators avoid manually recreating thousands of accounts.
• The transition is less disruptive before the on-premises data center is retired.

• Increase/decrease the number of epochs: Changes how long training runs, but does not reuse a pre-trained model.
• Unsupervised learning: Trains from unlabeled data and does not specifically address adapting an existing model.

• 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.

1. Define Power Platform DLP policies

1. Add custom prompt modifications

1. Conduct adversarial testing

• A: Using an external OAuth 2.0 identity provider could provide unified identity, but it does not specifically address Salesforce capacity or high-volume login limits.
• B: Making the Experience Cloud community a SAML identity provider and using JIT provisioning is a possible federation pattern, but it may not be appropriate for this scale without Salesforce validation.
• C: A default account is unrelated to solving identity unification and high login throughput. Person Accounts can be supported in appropriate B2C Commerce integrations.

• An Azure pay-as-you-go subscription follows an operational expenditure (OpEx) model.
• With OpEx, the organization pays for services as they are used, typically based on consumption.
• There is no need to purchase and maintain physical servers upfront.
• This contrasts with capital expenditure (CapEx), where the company makes a significant upfront investment in infrastructure, such as on-premises Hyper-V hosts.

• Customize Opportunity Contact Role

• Add the Opportunity Contact Role related list to the Opportunity page layout

• Campaign Member Role relates contacts or leads to campaigns, not opportunities.
• Campaign Member related list would expose campaign participation, not the stakeholders for the opportunity.
• Customize Campaign Role also concerns campaigns.

• A private IP address on the virtual network
• An associated public IP address

• A — Can help keep costs down and improve quality: The provider may have specialist expertise, efficient systems and economies of scale.
• B — There is a known fee if agreed in advance: A fixed or agreed fee can make costs more predictable and support budgeting.
• C — Allows management to focus on core operations and issues: Senior managers spend less time supervising routine finance processes and more time on strategic matters.

• Increased control: Control may decrease because the activity is performed externally.
• Increased independence: The organisation becomes dependent on the supplier.
• Increased understanding of the outsourcing business: The internal organisation usually has greater business knowledge; the provider may initially lack it.
• Increased confidentiality: Sharing financial information externally may create confidentiality risks.
• Internal expertise improved: Outsourcing can reduce opportunities to develop internal finance skills.

• The listener has a virtual IP address on the load balancer.
• The active SQL Server replica must respond using the listener’s IP address, not its own VM IP.
• Floating IP allows the listener IP to be preserved and returned by the active replica.
• This is required for the SQL Always On listener configuration.

• A health probe for the SQL listener/cluster role
• The SQL Server listener port, commonly 1433
• Appropriate backend pool membership

• VM CPU, memory, disk, and network usage
• Storage transactions and latency
• App Service response time
• Database DTU or CPU utilization

• Azure Traffic Analytics analyzes network traffic flows and security patterns; it is not the general infrastructure metrics tool.
• Azure Activity Log records subscription-level control-plane operations, such as resource creation, deletion, or configuration changes.
• Azure Advisor provides recommendations about cost, security, reliability, performance, and operational excellence, but it is not the primary tool for investigating live metrics.

• A — VARIANT "null" values are not the same as SQL NULL: Correct.

• B — Frequent, single-row DMLs: Incorrect.

• C — Validate data before loading into the target table: Correct.

• D — Use staging tables to manage MERGE statements: Correct.

• Workspace: Stores notebooks, libraries, jobs, and other project resources.
• Notebooks: Interactive documents containing code, queries, visualizations, and explanations. Common languages include Python, SQL, Scala, and R.
• Clusters: Groups of virtual machines that run Spark workloads. Clusters can be started, stopped, and automatically scaled.
• Data access: Databricks can read from services such as Azure Data Lake Storage, Azure Blob Storage, and Azure SQL.
• Jobs: Automate notebooks or data-processing workflows on a schedule or in response to events.

• Filtering or reshaping columns
• Joining with an analysis-specific dataset
• Creating calculated fields
• Applying business logic for a particular project

• Data blending: Combining different data sources, not specifically the final processing step.
• Data enhancement: A broad term for enriching data.
• Data testing: Checking data quality or pipeline correctness.
• Last-mile ETL: The precise term for additional transformations performed after the main pipeline, often directly before analytics or reporting.

• BCP: Maintains essential business operations during and after disruption. It is broader than IT recovery, but the question focuses on restoring the destroyed data center.
• Communication plan: Defines how the organization communicates with staff, the public, and other stakeholders.
• IRP: Handles security incidents such as malware, unauthorized access, or data breaches—not physical disaster recovery.

1. Identify every Region containing a replica.
2. Schedule deletion of the KMS key in each Region.
3. Use the minimum KMS waiting period: 7 days.

• A — Use the root user: Root access does not eliminate the need to schedule deletion for every regional replica. The problem is the multi-Region structure, not insufficient administrative permissions.

• C — Update the key policy: The policy already grants kms:* to IAM principals. Changing permissions does not delete the key or its replicas.

• D — Disable the key and schedule deletion in 30 days: Disabling can stop cryptographic use, but 30 days is slower than the permitted 7-day waiting period. It also does not address replicas in other Regions.

• A: Reduce management overhead — This is generally an advantage of public cloud, not hybrid cloud. Hybrid environments can increase management complexity.
• B: Cross-platform virtualization — Not the defining advantage of hybrid cloud.
• D: Leverage IaaS and PaaS — Public clouds commonly provide both IaaS and PaaS, so this is not unique to hybrid cloud.

• Equipment and hardware
• Facilities
• Inventory
• Backup media
• Supplies stored at an alternate or remote location

• Personal belongings of staff: Usually not organizational assets.
• DR line-item revenues: Revenue is a financial impact, not a physical asset.
• Cloud-based applications: These are primarily logical or information-processing assets, even though they depend on physical infrastructure.

• You normally do not create a new sensitive information type for credit card numbers. Microsoft Purview already provides a built-in Credit Card Number sensitive information type.
• The required sequence should generally be:

• Vulnerability scanner: Finds technical weaknesses in systems, applications, or devices. It does not usually analyze the entire supply chain.
• Penetration test: Simulates attacks against a defined target to identify exploitable weaknesses. Its scope is narrower.
• SCAP (Security Content Automation Protocol): Provides standardized formats and methods for automating security configuration and vulnerability checks, rather than mapping the full supply chain.

• Use one Azure AD directory for the organization.
• Separate divisions with subscriptions and/or resource groups.
• Assign each division’s administrator appropriate Azure role-based access control (RBAC) permissions, scoped to their subscription or resource group.

• 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.

• A: My Domain is required for many SSO configurations, but if SSO works, it is unlikely to be the cause of losing the target record.
• B: HTTP-POST is a valid SAML binding and does not inherently prevent redirection.
• D: An incorrect Federation ID generally causes authentication or user-matching failure—not successful login to the home page.

• Basic: No technical or architectural guidance.
• Standard: Technical support, but not the required design-assessment service.
• Professional Direct: Provides advisory and architecture guidance.
• Premier: Provides it too, but is unnecessarily expensive for this requirement.

• A Layer 2 switch checks the destination MAC address in its MAC address table (CAM table).
• If the destination MAC is unknown, the switch cannot select a specific outgoing port.
• It therefore performs unknown-unicast flooding, sending the frame out every other port in that VLAN.
• The frame is not sent back through the port where it arrived.

• 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.

• Amazon DynamoDB — Managed NoSQL database, not PostgreSQL.
• Amazon Athena — Serverless service for querying data in S3; intended for analytics, not OLTP.
• Amazon EMR — Managed cluster platform for big-data frameworks such as Hadoop and Spark.

• SAML: Enables federated authentication and single sign-on between the IdP and SaaS providers.
• OAuth tokens: Allow delegated application access; OAuth is not primarily the first step for employee SSO.
• Password vaulting: Stores or autofills multiple passwords but does not eliminate the separate credentials or provide true federation.

• Who performed the action
• What action occurred
• When it occurred
• Where it originated
• Whether it succeeded or failed

• Investigating suspicious activity: Important, but investigation depends on having adequate audit data first.
• Storage allocation: Necessary for retention, but storage does not help if the records lack meaningful details.
• Actions after audit failure: Part of operational planning, not the primary assessment of the application’s audit capability.

• Primary and core nodes: Use On-Demand Instances because the workload is business-critical. These nodes manage the cluster and store/process HDFS data.
• Task nodes: Use Spot Instances in an instance fleet. Task nodes are suitable for interruption because they do not hold essential cluster data and can be replaced or diversified across instance types.
• Terminate the entire cluster after processing: The job runs only about six hours each morning, and EMRFS stores data in Amazon S3. Keeping the cluster running for the rest of the day would waste money.

• A: Using Spot for primary and core nodes risks interruption of the critical cluster.
• C: All On-Demand is reliable but unnecessarily expensive; terminating the cluster helps, but Spot task nodes provide greater savings.
• D: Leaving primary and core nodes running after the task finishes continues to incur costs. Savings Plans also are less appropriate for this intermittent workload.

• The policy is confidential, and the requester is no longer authorized merely because they used to work for the organization.
• The security professional should not disclose it by email or display it on a device without authorization.
• Using official channels allows the organization’s owner or information custodian to verify: