CFA CFA I Exam Prep
CFA Level I Chartered Financial Analyst (Page 434 )

1. Search & Intelligence in the Microsoft 365 admin center — No

1. Audit in the Microsoft Defender portal — Yes

1. Audit in the Microsoft Purview portal — Yes

• C. An auto-labeling policy for a retention label
• D. A retention label

1. Create a retention label configured to:

1. Create an auto-labeling policy that detects the built-in Credit Card Number sensitive information type (SIT) and automatically applies the retention label to matching messages and documents.

• Data only needs to be current at 7 AM each day, so scheduled refresh is sufficient.
• Import mode provides the fastest report interaction because data is stored in Power BI’s in-memory engine.
• It supports row-level security (RLS), including the USERNAME() filter.
• You can limit the imported data to the current year and previous three years to reduce model size.
• DirectQuery would generally provide slower visual response times and is unnecessary here.
• Dual is mainly useful in composite models where some tables use DirectQuery and others use Import. That is not required by this scenario.
• Live connection is not appropriate because you are creating the semantic model.

• Transport and handling
• Storage and warehousing
• Refrigeration
• Shelf-stacking or replenishment
• Other costs directly attributable to product P

• Bronze: Raw ingested data
• Silver: Cleaned, validated, and transformed data
• Gold: Business-level data for analytics, dashboards, and reporting

1. Horizontal scaling — Yes

1. Vertical scaling — Yes

1. Scaling can manage costs — Yes

• Horizontal = more or fewer instances
• Vertical = larger or smaller instances

• Filtering harmful or inappropriate content
• Blocking specified topics
• Detecting and masking sensitive information
• Enforcing content policies suitable for children

• Amazon Rekognition analyzes images and videos; it does not control text generated by a language model.
• Amazon Bedrock playgrounds are for testing models and prompts, not enforcing production safety policies.
• Agents for Amazon Bedrock help models complete tasks and use tools; they do not primarily provide content-safety controls.

• Frequency/likelihood = how often or how likely the risk is to occur.
• Severity/impact = how damaging or costly the consequences would be.

• A minor equipment fault occurring frequently may have high frequency but low severity.
• A major cyberattack occurring rarely may have low frequency but high severity.
• A risk that happens often and causes substantial losses has high frequency and high severity.

• Sets a high initial price for a new product.
• Targets customers willing to pay more, often called early adopters.
• Gradually reduces the price to attract more price-sensitive customers as the product becomes less novel and competition increases.

• Penetration pricing: starts with a low price to gain market share quickly.
• Premium pricing: maintains a high price to communicate superior quality or status; it does not necessarily involve later price reductions.
• Loss leader pricing: sells one product at a very low price, sometimes below cost, to attract customers who may buy other products.

• Reduce the training data: Usually makes overfitting worse.
• Add hyperparameters: Hyperparameters already control training; simply adding them does not solve the problem. Proper tuning could help, but that is not what the option states.
• Increase training time: More training can worsen overfitting if the model is already memorizing the training data.

• A — Windows 11 with Azure CLI installed: Yes

• D — Chrome OS using Azure Cloud Shell: Yes

• E — Windows 10 with the Azure PowerShell module installed: Yes

• B — Linux with only Azure CLI: No, because the option does not state that PowerShell is installed.
• C — macOS with PowerShell Core only: No, because it does not state that Azure PowerShell modules or Azure CLI are installed.

• Malware or malicious software
• Command-and-control tools
• Internet surveillance and intrusion software

• ITAR: A U.S. regulation controlling defense-related articles and services; it is not the international arrangement described.
• Palermo Convention: Focuses on transnational organized crime and related protocols, not export controls for cyber weapons.
• GDPR: European data-protection and privacy legislation.

• Which batches and markets are affected
• Whether the deficiency affects product quality, efficacy, or patient safety
• Whether the issue is an ongoing manufacturing problem
• Whether immediate containment or distribution suspension is needed

• A — Withdraw the product: may be necessary, but is premature before assessing the risk.
• B — “Dear Dr.” letter: may be appropriate later, depending on the risk.
• C — Notify global authorities: may become mandatory, but the scope and urgency depend on the assessment.

• B. Versioned configuration files — Infrastructure is defined in machine-readable files and tracked in version control, allowing repeatable provisioning and change history.
• D. Docker files — Assuming “Doctor files” is a typo for Dockerfiles, a Dockerfile declaratively defines how to build a container image and is commonly considered infrastructure/configuration as code.

• A. Cloned virtual machine images — These are prebuilt infrastructure artifacts, but cloning an image is not itself a code-based infrastructure definition.
• C. Change management database records — A CMDB records information about infrastructure; it does not define or provision that infrastructure.

• GPS coordinates
• Date and time the image was taken
• Camera or device model
• Exposure settings

• Log data: Records events or activities, such as system or application activity.
• Encrypted data: Information intentionally encoded to prevent unauthorized access.
• Sensitive data: A broad classification for information requiring protection; geolocation may be sensitive, but the technical type stored in the image is metadata.

• Fine-tuning requires a properly formatted training dataset.
• For classification or instruction-style fine-tuning, examples are supplied in JSONL format.
• Each line represents one training example, typically containing the input prompt and the desired completion or label.
• The file must then be uploaded or made available to the Microsoft Foundry fine-tuning workflow.

• A: Prompt Flow is useful for prompt experimentation, but it is not required to start fine-tuning.
• B: You do not need to deploy the base model to an online endpoint before fine-tuning.
• C: Tracing supports observability and evaluation, but it does not prepare the fine-tuning job or dataset.

• FTP — Transfers files using a client/server model and supports authentication. It uses separate control and data connections.
• TFTP — A very simple file-transfer protocol with no authentication and minimal features. Commonly used to transfer router or switch configuration files and IOS images.
• SCP — Securely copies files over SSH. It provides encryption and authentication.
• SFTP — Also transfers files securely through SSH, but provides more file-management capabilities than traditional SCP.
• HTTP/HTTPS — Primarily web protocols, but can also download files; HTTPS encrypts the session.

• TFTP = simple, unauthenticated file transfer
• FTP = file transfer with authentication, but not inherently encrypted
• SCP/SFTP = secure file transfer using SSH

• Azure DevOps Analytics views are accessed through the Power BI Azure DevOps (Boards only) connector.
• This connector authenticates directly to Azure DevOps Services and lets you select an Analytics view.
• Analytics views allow filtering and shaping data in the cloud before Power BI loads it.
• Azure DevOps Server (Boards only) is for an on-premises Azure DevOps Server instance, not Azure DevOps cloud.
• A generic OData Feed connection can access OData endpoints, but it is not the purpose-built connector for Azure DevOps Analytics views.
• “OData queries” describes the underlying technology or query approach, not the most appropriate Azure DevOps connector choice here.

• Red team: Simulates attackers and tests exploitability.
• Blue team: Defends, monitors, detects, and responds to attacks.
• Purple team: Combines red- and blue-team activities to improve cooperation and defenses.
• White team: Manages and oversees the rules, scope, and authorization of the exercise.

• Data Cloud combines data from multiple systems.
• It resolves and organizes that data into a unified customer view.
• That unified view enables downstream actions such as personalized campaigns, segmentation, analytics, and activation.

• A = core platform value
• D = a resulting business use case or benefit

Step 1: Calculate the rights issue price

Step 2: Calculate the number of new shares

Step 3: Calculate the value after the issue and project

• Rights issue proceeds: $21.6m
• Project NPV: $5m

Step 4: Calculate the theoretical ex-rights price

• terraform init is normally required first, especially for a new or changed working directory. It installs providers/modules and configures the backend.
• terraform plan previews the proposed changes.
• terraform apply applies those changes. It also creates a plan internally if you have not supplied a saved plan.

text 
terraform init 
terraform validate       # optional but recommended 
terraform plan 
terraform apply

• The requirement is an assessment of the Azure environment’s design, which requires a support plan offering proactive architectural guidance.
• Premier support would provide this capability, but it is a high-cost option.
• Since the requirement says to minimize expenses, Premier is not the best recommendation.
• A lower-cost plan with the required advisory/design-review capability—historically Professional Direct in Azure’s support-plan model—would be more appropriate.

• Confirming the data has the required columns and data types
• Handling missing or duplicate records
• Converting data into a suitable format
• Selecting or engineering useful features
• Splitting it into training, validation, and test sets

1. Prepare the training data
2. Select an algorithm or configure AutoML
3. Train the model
4. Evaluate its performance
5. Tune hyperparameters if needed
6. Deploy the model

• Attack simulator — Used to run phishing and other attack simulations for user awareness training.
• Hunting — Used to proactively query security data and investigate potential threats.
• Incidents — Used to investigate and manage related alerts and active security events.

• Low: Applies system, tenant, and user-based suppressions. Activity from locations commonly used by that user is more likely to be suppressed.
• Medium: Applies system and user suppressions.
• High: Applies only system suppressions, so more unusual-location activity generates alerts.

• Active remediation actions role in Microsoft Defender for Endpoint

• Security Reader role in Azure AD/Entra ID

• Security Administrator is broader than necessary because it grants extensive security-management permissions, violating least privilege.
• Compliance Data Administrator is intended for compliance data management, not Defender for Endpoint remediation approvals.

• Instant messaging/chat: Can be abused to send malicious links or attachments, but P2P file sharing is the more direct high-risk source.
• E-mail: A common malware delivery method, but this question specifically emphasizes an application type associated with direct file exchange.
• End-to-end applications: This describes communication protection or architecture, not a specific application category known for malware distribution.

• Honeypots: Decoy systems designed to attract attackers. Legitimate users should not access them, so activity there is a strong indication of unauthorized access.
• DNS sinkholes: Redirect requests for known malicious domains to a controlled internal destination instead of the real attacker infrastructure. Security teams can then identify infected systems making those requests.

• Tokenization: Replaces sensitive data with non-sensitive tokens; it protects data but does not detect intruders.
• CI/CD: Software development and deployment processes.
• Threat modeling: Identifies potential threats before or during system design, rather than detecting successful infiltration.
• Data obfuscation: Makes information harder to understand, but is not an attacker-detection mechanism.

• Formatting: May remove file-system references but often leaves recoverable data.
• Degaussing: Uses a magnetic field to erase data, but can damage or make magnetic drives unusable.
• Defragmentation: Rearranges data for performance; it does not securely erase information.

• HTTPS: Secure web traffic, not general command-line server administration.
• SNMPv3: Secure network monitoring and management, not interactive server login.
• RDP: Provides graphical remote desktop access. It can be secure, but it is not the direct replacement for Telnet in this scenario.
• SMTP: Used to send email.

• Maintain head office control: Transfer pricing generally supports decentralisation rather than direct head-office control.
• Establish centralised decision making: This conflicts with divisional autonomy.
• Develop a top-down culture: Transfer pricing is intended to support empowered divisions and goal congruence, not impose purely top-down management.

• Honeypots are deliberately vulnerable or attractive systems designed to lure attackers. Legitimate users should not normally access them, so activity there can indicate a successful intrusion or lateral movement.

• A DNS sinkhole redirects requests for known malicious domains to a controlled internal system instead of the real destination. Security teams can monitor those requests and identify compromised devices attempting to contact command-and-control infrastructure.

• Tokenization: Replaces sensitive data with non-sensitive tokens; it is a protection method, not an attacker-detection method.
• CI/CD: Automates software development and deployment processes.
• Threat modeling: Identifies potential threats during design and planning.
• Data obfuscation: Makes information harder to understand, but does not directly detect infiltrators.

• Accept: Acknowledging the risk and choosing not to take additional action.
• Transfer: Shifting the risk or financial consequences to another party, such as an insurer.
• Mitigate: Reducing the likelihood or impact through controls, such as patching or access restrictions.
• Avoid: Eliminating the activity that creates the risk.

• A — Consumption-based pricing: AWS generally uses a pay-as-you-go model. You pay for the resources you consume rather than purchasing hardware upfront or paying for unused capacity.
• C — Economies of scale: AWS purchases and operates infrastructure at a very large scale, allowing it to reduce costs and pass some savings to customers through lower service prices.

• Perpetual licenses: These involve paying once for permanent software ownership and are not a primary AWS Cloud economic benefit.
• AWS Enterprise Support at no additional cost: Enterprise Support has an additional charge.
• Bring-your-own-hardware model: AWS provides and manages the underlying cloud infrastructure; customers do not normally bring their own physical servers.

• IRR assumes that interim cash flows can be reinvested at the project’s own IRR. This may be unrealistic, especially when the IRR is very high.
• MIRR assumes reinvestment at a specified rate, usually the company’s cost of capital or another realistic reinvestment rate.

• B: MIRR does not specifically favour projects with long payback periods.
• C: MIRR and IRR can rank competing projects differently.
• D: MIRR is not always higher than IRR; it may be higher or lower depending on the cash flows and reinvestment rate.

• Run the backup with WITH CHECKSUM
• Store the backup on an SMB network share

sql 
BACKUP DATABASE DB1 
TO DISK = '\\FileServer\SqlBackups\DB1_full.bak' 
WITH CHECKSUM;

• DMS does not create the SQL Server backups for you.
• You must provide a valid full database backup.
• You must continue taking transaction log backups while the source database remains online.
• The full backup and subsequent log backups must be available in the SMB share that DMS can access.
• WITH CHECKSUM enables backup checksum validation, helping detect backup corruption.

• SAS (Serial Attached SCSI):

• SSD (solid-state drive):

• Input validation: Prevents malicious or malformed input within an application; it does not authenticate a software package.
• Secure cookies: Protect web-session cookies; they do not protect downloadable software.
• Fuzzing: Tests software by supplying unexpected input to find vulnerabilities; it does not verify software integrity.

sql 
ALTER TABLE dbo.DimCustomer 
ADD CONSTRAINT PK_DimCustomer 
PRIMARY KEY NONCLUSTERED (CustomerKey) 
NOT ENFORCED;

1. ALTER TABLE dbo.DimCustomer
2. ADD CONSTRAINT PK_DimCustomer PRIMARY KEY NONCLUSTERED (CustomerKey)
3. NOT ENFORCED

• A single summarized value, such as total annual sales
• Progress toward a defined target or goal
• Whether performance is below, near, or above the target

• Table: Displays detailed records or summarized rows, but not goal progress visually.
• Stacked Bar: Compares multiple categories and their parts.
• Donut: Shows how a total is divided among categories, not progress toward a target.

• sn_si_admin is an SIR administrator role.
• The prefix helps prevent naming conflicts between applications.
• It does not refer to a “Secure Identity token” or an instance security check.

• A is correct: It establishes the revised dates and identifies the effect on the critical path.
• B is premature: Stakeholders should be informed after the impact has been assessed and the updated schedule is understood.
• C is not automatically appropriate: Fast-tracking may introduce additional risk and should be considered only after analyzing schedule options.
• D is premature: Adding resources may not resolve a dependency and could increase costs without addressing the actual constraint.

• Response and resolution time frames
• Availability or uptime requirements
• Performance targets
• Support obligations
• Penalties for failing to meet requirements

• SOW (Statement of Work): Defines the project’s scope, tasks, and deliverables.
• MOA (Memorandum of Agreement): Describes cooperation and responsibilities between parties.
• MOU (Memorandum of Understanding): Records a general mutual understanding, usually without detailed enforceable service targets.

• Labels control access and protection, not blanket exclusion.
• If the user has permission to open a labeled file—especially if the label allows authenticated users to view it—Copilot can use that file in a summary.
• If the label applies encryption and the user lacks the required usage rights, Copilot should not use the file or reveal its contents.
• Copilot honors the user’s existing permissions, so it should not expose labeled content to an unauthorized user.

• Domain controllers, because the requirement specifically excludes them. However, note that both member servers and domain controllers run Windows Server, so OS targeting alone cannot distinguish them unless the targeting option allows selecting the relevant server role/version. In the usual exam interpretation, target the applicable server OS category.
• Client computers, because they run a client Windows operating system.

• Domain: Targets membership in a domain, not the computer’s operating-system type.
• Security Group: Would require creating and maintaining a group containing all member servers, increasing administrative effort.
• Environment Variable: Not designed to reliably identify member servers.

• VLAN 10 for data
• VLAN 20 for voice

• A: Describes physical trunking or logical switch ports, not an SVI.
• B: An SVI can provide management access, but that is only one use—not the best description here.
• D: SVI use is not related to running out of physical ports.

• Tokens: Individual units of text processed by a model, such as words, subwords, or characters.
• Models: Systems that learn patterns and generate predictions or outputs.
• Binaries: Data represented using the binary number system, not semantic representations.

• The JSON files contain one billion items, so this is a big-data workload.
• PySpark uses Apache Spark’s distributed, parallel processing, allowing the data to be processed across multiple compute nodes.
• It can read and transform data directly from OneLake, helping minimize unnecessary data movement or duplication.
• A Fabric notebook can also support exploration and visualization, as well as Spark-based analytics and anomaly-detection workflows.

• Pandas is primarily in-memory and generally suited to smaller datasets. One billion records would be impractical on a typical single machine.
• A Power BI report with core visuals is appropriate for visualization and sharing, but it is not the best tool for large-scale transformation, parallel processing, or anomaly detection. It would normally consume prepared data rather than perform the entire workload.

• Microsoft Foundry is an enterprise platform for developers, data scientists, and IT teams.
• It supports the full AI application lifecycle: selecting models, connecting enterprise data, developing, evaluating, deploying, monitoring, and scaling solutions.
• It can include visual tools and low-code experiences, but it is not primarily designed to enable business users to build complete generative AI solutions independently.

• B: You still need to select and configure an appropriate model.
• C: Business users may participate through visual tools, but they are not the primary target audience.
• D: Foundry includes low-code tools, but it is not fundamentally a low-code-only platform.