Topic 02 Practice Test 1 covers Multimodal Input Validation, Processing, Formatting, and Data Quality in Domain 1 (Task 1.3). For broader exam preparation, review the AIP-C01 Exam Dumps. Every option includes focused technical reasoning tied to the tested decision.
Question 1
A developer is hardening an FM-consumption workflow. A nightly S3 dataset sometimes omits the customer_intent field. The FM must never receive rows where that required field is null. Which approach most directly addresses the stated constraint? Choose ONE.
- checksum integrity verification
- AWS Glue Data Quality completeness rule
- Lambda JSON Schema validation
- required-field freshness validation
Correct Answer: B
Correct Answer
Answer B is correct because AWS Glue Data Quality completeness rule is the best fit because it can evaluate required columns for nulls before records enter the FM pipeline. The workflow needs a case where a batch table has mandatory fields and incomplete rows must be rejected.
Incorrect Answers
Answer A is incorrect because checksum integrity verification can help a pipeline compare a supplied cryptographic checksum with the downloaded object before preprocessing, especially when large artifacts can be corrupted in transit or during transfer and must be verified before use. The present requirement is a batch table has mandatory fields and incomplete rows must be rejected, so it is not the best answer.
Answer C is incorrect because Lambda JSON Schema validation works for cases where small event payloads need custom application-level validation at ingestion time by helping to enforce the expected structure and data types for each event before invoking the model. The present scenario needs a batch table has mandatory fields and incomplete rows must be rejected.
Answer D is incorrect because required-field freshness validation targets a neighboring concern. It can compare a source timestamp with an allowed staleness window before accepting the record when the FM must operate only on sufficiently recent operational data. The key requirement here is a batch table has mandatory fields and incomplete rows must be rejected.
Question 2
An application team is standardizing data before foundation-model inference. A telemetry table contains temperature_c values. Any value below -80 or above 80 must be quarantined before it is summarized. Which approach most directly addresses the stated constraint? Choose ONE.
- S3 quarantine prefix for failed validation
- AWS Glue Data Quality allowed-value rule
- AWS Glue Data Quality numeric-range rule
- checksum integrity verification
Correct Answer: C
Correct Answer
Answer C is correct because AWS Glue Data Quality numeric-range rule is appropriate because it will check numeric values against expected minimum and maximum boundaries as part of batch quality validation. The requirement centers on a situation where out-of-range measurements would mislead the FM.
Incorrect Answers
Answer A is incorrect because Using S3 quarantine prefix for failed validation would let the team isolate records that fail validation from the clean processing path while retaining them for investigation. That is appropriate when invalid objects must be retained but prevented from reaching FM consumption. The scenario instead calls for out-of-range measurements would mislead the FM.
Answer B is incorrect because AWS Glue Data Quality allowed-value rule has value because it can validate a categorical column against an approved domain before downstream processing. Its proper fit is when only enumerated labels are valid for a field consumed by prompting logic. This question requires out-of-range measurements would mislead the FM.
Answer D is incorrect because checksum integrity verification would improve another stage by enabling it to compare a supplied cryptographic checksum with the downloaded object before preprocessing. It fits when large artifacts can be corrupted in transit or during transfer and must be verified before use. The stated problem is out-of-range measurements would mislead the FM.
Question 3
A platform team is reviewing a production GenAI data pipeline. Users upload .png files, but some are actually renamed executables or corrupted data. The pipeline must verify the file type before model use. Which approach most directly addresses the stated constraint? Choose ONE.
- AWS Glue Data Quality allowed-value rule
- tabular type coercion with explicit failure handling
- AWS Glue Data Quality completeness rule
- Lambda file-signature validation
Correct Answer: D
Correct Answer
Answer D is correct because Select Lambda file-signature validation. Its purpose is to inspect magic bytes and declared content type so mislabeled or unsupported files are rejected early. The operational requirement is that uploaded files can have misleading extensions and must be validated before multimodal processing.
Incorrect Answers
Answer A is incorrect because AWS Glue Data Quality allowed-value rule is useful when only enumerated labels are valid for a field consumed by prompting logic because it can validate a categorical column against an approved domain before downstream processing. The current requirement is uploaded files can have misleading extensions and must be validated before multimodal processing, which makes this a mismatch.
Answer B is incorrect because tabular type coercion with explicit failure handling addresses another concern by helping to convert fields to expected types and route conversion failures to an error path instead of silently guessing. It fits where structured inputs contain type drift that must be made deterministic before FM formatting. Here the deciding need is uploaded files can have misleading extensions and must be validated before multimodal processing.
Answer C is incorrect because Using AWS Glue Data Quality completeness rule would let the team evaluate required columns for nulls before records enter the FM pipeline. That is appropriate when a batch table has mandatory fields and incomplete rows must be rejected. The scenario instead calls for uploaded files can have misleading extensions and must be validated before multimodal processing.
Question 4
A cloud engineering team must correct a data-processing weakness. A partner provides a SHA-256 checksum for each training-reference archive uploaded to S3. The pipeline must detect silent corruption before extraction. Which approach most directly addresses the stated constraint? Choose ONE.
- S3 quarantine prefix for failed validation
- SageMaker Data Wrangler profiling
- checksum integrity verification
- deduplication by stable record identifier
Correct Answer: C
Correct Answer
Answer C is correct because checksum integrity verification is the direct remedy. It can compare a supplied cryptographic checksum with the downloaded object before preprocessing. The critical scenario condition is that large artifacts can be corrupted in transit or during transfer and must be verified before use.
Incorrect Answers
Answer A is incorrect because S3 quarantine prefix for failed validation targets a different issue by helping to isolate records that fail validation from the clean processing path while retaining them for investigation. It fits where invalid objects must be retained but prevented from reaching FM consumption. The deciding condition is large artifacts can be corrupted in transit or during transfer and must be verified before use.
Answer B is incorrect because SageMaker Data Wrangler profiling can profile distributions, missingness, and anomalies interactively before defining production transformations. It fits when a team must understand data quality problems before finalizing preprocessing logic. This scenario instead requires large artifacts can be corrupted in transit or during transfer and must be verified before use, so it targets a different problem.
Answer D is incorrect because deduplication by stable record identifier addresses another concern by helping to drop repeated records using a durable business key before constructing FM context. It fits where retries create duplicate records that would otherwise overweight the same evidence. Here the deciding need is large artifacts can be corrupted in transit or during transfer and must be verified before use.
Question 5
A regulated workflow requires deterministic preprocessing before inference. An upstream webhook can deliver the same ticket event multiple times with the same event_id. The summarizer should consider each event once. Which approach most directly addresses the stated constraint? Choose ONE.
- deduplication by stable record identifier
- model input-size precheck
- AWS Glue Data Quality uniqueness rule
- checksum integrity verification
Correct Answer: A
Correct Answer
Answer A is correct because deduplication by stable record identifier is the best fit because it can drop repeated records using a durable business key before constructing FM context. The workflow needs a case where retries create duplicate records that would otherwise overweight the same evidence.
Incorrect Answers
Answer B is incorrect because model input-size precheck is useful when requests exceeding model limits should be handled deterministically before calling the FM and can estimate payload size or token count before invocation and reject or route oversized input. The described workload requires retries create duplicate records that would otherwise overweight the same evidence.
Answer C is incorrect because AWS Glue Data Quality uniqueness rule targets a different issue by helping to measure duplicate keys in a managed ruleset and fail the quality check when uniqueness drops. It fits where duplicate business keys in a tabular batch would cause repeated evidence. The deciding condition is retries create duplicate records that would otherwise overweight the same evidence.
Answer D is incorrect because checksum integrity verification can compare a supplied cryptographic checksum with the downloaded object before preprocessing. It fits when large artifacts can be corrupted in transit or during transfer and must be verified before use. This scenario instead requires retries create duplicate records that would otherwise overweight the same evidence, so it targets a different problem.
Question 6
A developer is hardening an FM-consumption workflow. A call-center application stores WAV recordings and needs a textual transcript before the FM generates a case summary. Which approach most directly addresses the stated constraint? Choose ONE.
- Amazon Transcribe speech-to-text preprocessing
- image dimension and format normalization
- parallel modality preprocessing
- table-header preservation during extraction
Correct Answer: A
Correct Answer
Answer A is correct because Amazon Transcribe speech-to-text preprocessing is appropriate because it will convert spoken audio into text with timestamps before text-oriented FM processing. The requirement centers on a situation where audio must become searchable and summarizable text before a text-only downstream step.
Incorrect Answers
Answer B is incorrect because image dimension and format normalization can resize oversized images and convert them to a supported format before model invocation, which matters when input images exceed model/service constraints even though their visual content is valid. The present case is driven by audio must become searchable and summarizable text before a text-only downstream step.
Answer C is incorrect because parallel modality preprocessing belongs in workflows where multiple independent modalities create unnecessary latency when handled serially because it can process independent audio, image, and text components concurrently before joining them for inference. This question instead asks for audio must become searchable and summarizable text before a text-only downstream step.
Answer D is incorrect because table-header preservation during extraction is useful when separating values from their headers would destroy meaning needed by the FM and can retain column headers with row values when converting tables to text or structured records. The described workload requires audio must become searchable and summarizable text before a text-only downstream step.
Question 7
An application team is standardizing data before foundation-model inference. A claims system receives scanned forms with boxes, labels, and tables. The FM needs the extracted fields and table content. Which approach most directly addresses the stated constraint? Choose ONE.
- table-header preservation during extraction
- Amazon Textract document extraction
- tabular rows serialized as structured JSON
- SageMaker Processing batch media preprocessing
Correct Answer: B
Correct Answer
Answer B is correct because Select Amazon Textract document extraction. Its purpose is to extract printed text, forms, and table structure from scanned documents before downstream language processing. The operational requirement is that scanned forms need structured text and key-value information before an FM can reason over them.
Incorrect Answers
Answer A is incorrect because table-header preservation during extraction solves a different input problem. It helps to retain column headers with row values when converting tables to text or structured records, normally when separating values from their headers would destroy meaning needed by the FM. The current workflow needs scanned forms need structured text and key-value information before an FM can reason over them.
Answer C is incorrect because tabular rows serialized as structured JSON would be sound if the goal were a tabular record must be represented clearly for inference rather than as an unlabeled value sequence; it can convert selected columns into explicit key-value objects so the FM receives unambiguous field names and values. The requirement here is scanned forms need structured text and key-value information before an FM can reason over them.
Answer D is incorrect because SageMaker Processing batch media preprocessing can run scalable, repeatable preprocessing jobs over large media batches without tying work to request-time compute, which matters when thousands of files need scheduled transformation before FM consumption. The present case is driven by scanned forms need structured text and key-value information before an FM can reason over them.
Question 8
A platform team is reviewing a production GenAI data pipeline. A pricing assistant receives rows from a relational table and must preserve the meaning of sku, region, currency, and amount in each prompt. Which approach most directly addresses the stated constraint? Choose ONE.
- Step Functions multimodal orchestration
- tabular rows serialized as structured JSON
- low-confidence OCR routing to human review
- event-driven S3 Lambda preprocessing
Correct Answer: B
Correct Answer
Answer B is correct because tabular rows serialized as structured JSON is the direct remedy. It can convert selected columns into explicit key-value objects so the FM receives unambiguous field names and values. The critical scenario condition is that a tabular record must be represented clearly for inference rather than as an unlabeled value sequence.
Incorrect Answers
Answer A is incorrect because Step Functions multimodal orchestration helps when a multi-stage preprocessing workflow needs durable state and error handling by allowing the pipeline to coordinate separate extraction, normalization, validation, and FM steps with explicit retries and failure branches. The described issue is a tabular record must be represented clearly for inference rather than as an unlabeled value sequence.
Answer C is incorrect because low-confidence OCR routing to human review solves a different input problem. It helps to branch documents with unreliable extraction scores to manual verification before FM use, normally when uncertain extracted text must not silently become authoritative model context. The current workflow needs a tabular record must be represented clearly for inference rather than as an unlabeled value sequence.
Answer D is incorrect because event-driven S3 Lambda preprocessing is appropriate where small independent objects should begin processing immediately without a polling batch job. It achieves that by helping to trigger lightweight preprocessing automatically when a new object lands in S3. This scenario instead requires a tabular record must be represented clearly for inference rather than as an unlabeled value sequence.
Question 9
A cloud engineering team must correct a data-processing weakness. A support case contains a screenshot, an audio note, and typed text. Each can be preprocessed independently before the final multimodal prompt. Which approach most directly addresses the stated constraint? Choose ONE.
- video key-frame extraction plus transcript pipeline
- Amazon Textract document extraction
- parallel modality preprocessing
- language-aware transcription routing
Correct Answer: C
Correct Answer
Answer C is correct because parallel modality preprocessing is the best fit because it can process independent audio, image, and text components concurrently before joining them for inference. The workflow needs a case where multiple independent modalities create unnecessary latency when handled serially.
Incorrect Answers
Answer A is incorrect because video key-frame extraction plus transcript pipeline is designed to reduce a video to representative visual frames and speech text before multimodal reasoning. That makes sense when full video is too large for direct use but both visual and spoken evidence matter. The question is centered on multiple independent modalities create unnecessary latency when handled serially.
Answer B is incorrect because Amazon Textract document extraction can extract printed text, forms, and table structure from scanned documents before downstream language processing. Its best fit is where scanned forms need structured text and key-value information before an FM can reason over them. The current case instead needs multiple independent modalities create unnecessary latency when handled serially.
Answer D is incorrect because language-aware transcription routing is relevant, but it is intended to detect or configure the spoken language so the speech service uses the appropriate recognition path when audio arrives in multiple languages and a single hard-coded language causes poor transcripts. The actual requirement is multiple independent modalities create unnecessary latency when handled serially.
Question 10
A regulated workflow requires deterministic preprocessing before inference. A 12-GB log export must be cleaned and transformed for later FM analysis without loading the entire file into memory. Which approach most directly addresses the stated constraint? Choose ONE.
- tabular rows serialized as structured JSON
- image dimension and format normalization
- video key-frame extraction plus transcript pipeline
- large-file chunked preprocessing
Correct Answer: D
Correct Answer
Answer D is correct because large-file chunked preprocessing is appropriate because it will split a very large source artifact into bounded processing units before normalization and later reassembly. The requirement centers on a situation where a single preprocessing job would exceed memory or request limits.
Incorrect Answers
Answer A is incorrect because tabular rows serialized as structured JSON works for cases where a tabular record must be represented clearly for inference rather than as an unlabeled value sequence by helping to convert selected columns into explicit key-value objects so the FM receives unambiguous field names and values. The present scenario needs a single preprocessing job would exceed memory or request limits.
Answer B is incorrect because image dimension and format normalization targets a neighboring concern. It can resize oversized images and convert them to a supported format before model invocation when input images exceed model/service constraints even though their visual content is valid. The key requirement here is a single preprocessing job would exceed memory or request limits.
Answer C is incorrect because video key-frame extraction plus transcript pipeline is designed to reduce a video to representative visual frames and speech text before multimodal reasoning. That makes sense when full video is too large for direct use but both visual and spoken evidence matter. The question is centered on a single preprocessing job would exceed memory or request limits.
Question 11
A developer is hardening an FM-consumption workflow. An application switches from one Bedrock model family to another and receives validation errors because it reuses the previous provider’s request fields. Which approach most directly addresses the stated constraint? Choose ONE.
- Bedrock Converse message format
- explicit JSON numeric and Boolean types
- model-specific Bedrock request serialization
- ordered conversation-turn serialization
Correct Answer: C
Correct Answer
Answer C is correct because Select model-specific Bedrock request serialization. Its purpose is to build the JSON body expected by the selected model instead of assuming every provider uses the same native schema. The operational requirement is that native InvokeModel payloads differ among model providers.
Incorrect Answers
Answer A is incorrect because Bedrock Converse message format can help a pipeline represent dialog turns with the standardized role/content message structure used by the Converse API, especially when a multi-turn application needs a consistent conversational request format across supported models. The present requirement is native InvokeModel payloads differ among model providers, so it is not the best answer.
Answer B is incorrect because explicit JSON numeric and Boolean types would improve another stage by enabling it to serialize values using the data types required by the model request schema rather than converting everything to strings. It fits when schema-valid structured request fields must retain their native types. The stated problem is native InvokeModel payloads differ among model providers.
Answer D is incorrect because ordered conversation-turn serialization targets a neighboring concern. It can preserve chronological role order when converting stored chat history into the model request when reordered dialog turns would change the meaning of follow-up questions. The key requirement here is native InvokeModel payloads differ among model providers.
Question 12
An application team is standardizing data before foundation-model inference. A multimodal request fails because PNG bytes are inserted into a text field without declaring image content and its media type. Which approach most directly addresses the stated constraint? Choose ONE.
- speaker-label mapping to conversation roles
- ordered conversation-turn serialization
- base64 or binary image content with correct media type
- system instruction separated from user content
Correct Answer: C
Correct Answer
Answer C is correct because base64 or binary image content with correct media type is the direct remedy. It can encode image content using the request format and MIME type expected by the multimodal API. The critical scenario condition is that visual input is valid but the request currently treats binary bytes as ordinary text.
Incorrect Answers
Answer A is incorrect because Using speaker-label mapping to conversation roles would let the team map source labels such as AGENT and CUSTOMER into the model API’s supported conversational roles. That is appropriate when source-system speaker names are not directly valid message roles for the inference API. The scenario instead calls for visual input is valid but the request currently treats binary bytes as ordinary text.
Answer B is incorrect because ordered conversation-turn serialization has value because it can preserve chronological role order when converting stored chat history into the model request. Its proper fit is when reordered dialog turns would change the meaning of follow-up questions. This question requires visual input is valid but the request currently treats binary bytes as ordinary text.
Answer D is incorrect because system instruction separated from user content would improve another stage by enabling it to place persistent behavioral instructions in the system field while keeping the user’s actual request in the user message. It fits when the application must clearly distinguish global behavior from per-turn input. The stated problem is visual input is valid but the request currently treats binary bytes as ordinary text.
Question 13
A platform team is reviewing a production GenAI data pipeline. A Bedrock request fails validation because maxTokens is serialized as the string ‘512’ and stream is serialized as the string ‘false’. Which approach most directly addresses the stated constraint? Choose ONE.
- explicit JSON numeric and Boolean types
- provider-versioned request adapter
- model-specific Bedrock request serialization
- conversation attachment content blocks
Correct Answer: A
Correct Answer
Answer A is correct because explicit JSON numeric and Boolean types is the best fit because it can serialize values using the data types required by the model request schema rather than converting everything to strings. The workflow needs a case where schema-valid structured request fields must retain their native types.
Incorrect Answers
Answer B is incorrect because provider-versioned request adapter addresses another concern by helping to route each selected model through an adapter that owns that model family’s request and response mapping. It fits where an application supports multiple model families whose native schemas evolve independently. Here the deciding need is schema-valid structured request fields must retain their native types.
Answer C is incorrect because Using model-specific Bedrock request serialization would let the team build the JSON body expected by the selected model instead of assuming every provider uses the same native schema. That is appropriate when native InvokeModel payloads differ among model providers. The scenario instead calls for schema-valid structured request fields must retain their native types.
Answer D is incorrect because conversation attachment content blocks has value because it can represent text and supported attachments as separate typed content blocks within one conversational message. Its proper fit is when a message combines written instructions with a document or image and each part must retain its type. This question requires schema-valid structured request fields must retain their native types.
Question 14
A cloud engineering team must correct a data-processing weakness. The request payload is valid JSON, but the endpoint rejects it because the client sends text/plain rather than the required JSON content type. Which approach most directly addresses the stated constraint? Choose ONE.
- correct HTTP Content-Type for JSON inference
- batch-inference record envelope
- system instruction separated from user content
- explicit JSON numeric and Boolean types
Correct Answer: A
Correct Answer
Answer A is correct because correct HTTP Content-Type for JSON inference is appropriate because it will send the media type expected by the inference endpoint together with the serialized request body. The requirement centers on a situation where a syntactically valid JSON body is rejected because the protocol metadata is wrong.
Incorrect Answers
Answer B is incorrect because batch-inference record envelope can emit each batch item using the service’s required record identifier and model-input envelope. It fits when offline inference requires per-record association so outputs can be mapped back to source items. This scenario instead requires a syntactically valid JSON body is rejected because the protocol metadata is wrong, so it targets a different problem.
Answer C is incorrect because system instruction separated from user content is useful when the application must clearly distinguish global behavior from per-turn input because it can place persistent behavioral instructions in the system field while keeping the user’s actual request in the user message. The current requirement is a syntactically valid JSON body is rejected because the protocol metadata is wrong, which makes this a mismatch.
Answer D is incorrect because explicit JSON numeric and Boolean types addresses another concern by helping to serialize values using the data types required by the model request schema rather than converting everything to strings. It fits where schema-valid structured request fields must retain their native types. Here the deciding need is a syntactically valid JSON body is rejected because the protocol metadata is wrong.
Question 15
A regulated workflow requires deterministic preprocessing before inference. A nightly batch job submits thousands of model requests and must reliably correlate every generated response with the originating record. Which approach most directly addresses the stated constraint? Choose ONE.
- explicit JSON numeric and Boolean types
- normalized UTF-8 text encoding
- speaker-label mapping to conversation roles
- batch-inference record envelope
Correct Answer: D
Correct Answer
Answer D is correct because Select batch-inference record envelope. Its purpose is to emit each batch item using the service’s required record identifier and model-input envelope. The operational requirement is that offline inference requires per-record association so outputs can be mapped back to source items.
Incorrect Answers
Answer A is incorrect because explicit JSON numeric and Boolean types belongs in workflows where schema-valid structured request fields must retain their native types because it can serialize values using the data types required by the model request schema rather than converting everything to strings. This question instead asks for offline inference requires per-record association so outputs can be mapped back to source items.
Answer B is incorrect because normalized UTF-8 text encoding is useful when mixed encodings create invalid characters and inconsistent tokenization and can convert incoming text to a consistent Unicode encoding before request serialization. The described workload requires offline inference requires per-record association so outputs can be mapped back to source items.
Answer C is incorrect because speaker-label mapping to conversation roles targets a different issue by helping to map source labels such as AGENT and CUSTOMER into the model API’s supported conversational roles. It fits where source-system speaker names are not directly valid message roles for the inference API. The deciding condition is offline inference requires per-record association so outputs can be mapped back to source items.
Question 16
A developer is hardening an FM-consumption workflow. A document workflow needs to tag company and location names before the FM receives the normalized text and metadata. Which approach most directly addresses the stated constraint? Choose ONE.
- Amazon Comprehend entity extraction
- Unicode normalization
- Bedrock-assisted text reformatting with constrained output
- exact-content deduplication
Correct Answer: A
Correct Answer
Answer A is correct because Amazon Comprehend entity extraction is the direct remedy. It can identify entities such as organizations, people, and locations before enriching the model input. The critical scenario condition is that explicit entity metadata will improve downstream context and routing.
Incorrect Answers
Answer B is incorrect because Unicode normalization can convert canonically equivalent Unicode sequences to one representation before deduplication and inference, which matters when visually identical text is represented by different code-point sequences. The present case is driven by explicit entity metadata will improve downstream context and routing.
Answer C is incorrect because Bedrock-assisted text reformatting with constrained output belongs in workflows where source text is semistructured and deterministic parsing alone cannot reliably produce the required representation because it can use an FM preprocessing step to transform irregular prose into a defined structure while validating the result afterward. This question instead asks for explicit entity metadata will improve downstream context and routing.
Answer D is incorrect because exact-content deduplication is useful when duplicated passages waste tokens and can overweight one source and can remove repeated identical text blocks before assembling the model context. The described workload requires explicit entity metadata will improve downstream context and routing.
Question 17
An application team is standardizing data before foundation-model inference. Records contain 03/04/2026, 2026-04-03, and ‘April 3 2026’. The FM input should represent dates consistently and unambiguously. Which approach most directly addresses the stated constraint? Choose ONE.
- Unicode normalization
- entity metadata enrichment
- whitespace normalization
- date-format canonicalization
Correct Answer: D
Correct Answer
Answer D is correct because date-format canonicalization is the best fit because it can convert multiple source date conventions into one explicit machine-readable format before model consumption. The workflow needs a case where ambiguous date strings can be interpreted differently across locales.
Incorrect Answers
Answer A is incorrect because Unicode normalization solves a different input problem. It helps to convert canonically equivalent Unicode sequences to one representation before deduplication and inference, normally when visually identical text is represented by different code-point sequences. The current workflow needs ambiguous date strings can be interpreted differently across locales.
Answer B is incorrect because entity metadata enrichment is appropriate where the text alone is usable but extra structured metadata improves downstream selection. It achieves that by helping to attach extracted entity attributes to the normalized record so later prompts can filter or contextualize content. This scenario instead requires ambiguous date strings can be interpreted differently across locales.
Answer C is incorrect because whitespace normalization would be sound if the goal were formatting noise creates unnecessary tokens without carrying meaning; it can collapse accidental repeated spaces and normalize line endings while preserving semantic paragraph boundaries. The requirement here is ambiguous date strings can be interpreted differently across locales.
Question 18
A platform team is reviewing a production GenAI data pipeline. An internal crawler captures the same sidebar links and legal footer on every page before the text is sent to the FM. Which approach most directly addresses the stated constraint? Choose ONE.
- date-format canonicalization
- navigation-boilerplate removal
- speaker-turn cleanup and labeling
- exact-content deduplication
Correct Answer: B
Correct Answer
Answer B is correct because navigation-boilerplate removal is appropriate because it will strip repeated menus, cookie notices, and site chrome while retaining article content. The requirement centers on a situation where web-page boilerplate consumes context and distracts from the source information.
Incorrect Answers
Answer A is incorrect because date-format canonicalization helps when ambiguous date strings can be interpreted differently across locales by allowing the pipeline to convert multiple source date conventions into one explicit machine-readable format before model consumption. The described issue is web-page boilerplate consumes context and distracts from the source information.
Answer C is incorrect because speaker-turn cleanup and labeling solves a different input problem. It helps to remove transcription artifacts and preserve who said each utterance before passing a conversation to the FM, normally when a transcript contains repeated filler markers and ambiguous speaker boundaries. The current workflow needs web-page boilerplate consumes context and distracts from the source information.
Answer D is incorrect because exact-content deduplication is appropriate where duplicated passages waste tokens and can overweight one source. It achieves that by helping to remove repeated identical text blocks before assembling the model context. This scenario instead requires web-page boilerplate consumes context and distracts from the source information.
Question 19
A cloud engineering team must correct a data-processing weakness. A global support corpus contains French, German, and Japanese notes. The workflow must identify the language before selecting the correct prompt and model path. Which approach most directly addresses the stated constraint? Choose ONE.
- Unicode normalization
- language detection and language-specific routing
- whitespace normalization
- Amazon Comprehend entity extraction
Correct Answer: B
Correct Answer
Answer B is correct because Select language detection and language-specific routing. Its purpose is to detect the dominant language and route the record to a language-appropriate prompt or supported model path. The operational requirement is that documents arrive in several languages and the pipeline must select processing that matches each language.
Incorrect Answers
Answer A is incorrect because Unicode normalization is designed to convert canonically equivalent Unicode sequences to one representation before deduplication and inference. That makes sense when visually identical text is represented by different code-point sequences. The question is centered on documents arrive in several languages and the pipeline must select processing that matches each language.
Answer C is incorrect because whitespace normalization helps when formatting noise creates unnecessary tokens without carrying meaning by allowing the pipeline to collapse accidental repeated spaces and normalize line endings while preserving semantic paragraph boundaries. The described issue is documents arrive in several languages and the pipeline must select processing that matches each language.
Answer D is incorrect because Amazon Comprehend entity extraction is relevant, but it is intended to identify entities such as organizations, people, and locations before enriching the model input when explicit entity metadata will improve downstream context and routing. The actual requirement is documents arrive in several languages and the pipeline must select processing that matches each language.
Question 20
A regulated workflow requires deterministic preprocessing before inference. Contracts should be enriched with detected organization and location metadata so later workflows can select only documents relevant to a specific customer. Which approach most directly addresses the stated constraint? Choose ONE.
- Unicode normalization
- Bedrock-assisted text reformatting with constrained output
- exact-content deduplication
- entity metadata enrichment
Correct Answer: D
Correct Answer
Answer D is correct because entity metadata enrichment is the direct remedy. It can attach extracted entity attributes to the normalized record so later prompts can filter or contextualize content. The critical scenario condition is that the text alone is usable but extra structured metadata improves downstream selection.
Incorrect Answers
Answer A is incorrect because Unicode normalization works for cases where visually identical text is represented by different code-point sequences by helping to convert canonically equivalent Unicode sequences to one representation before deduplication and inference. The present scenario needs the text alone is usable but extra structured metadata improves downstream selection.
Answer B is incorrect because Bedrock-assisted text reformatting with constrained output targets a neighboring concern. It can use an FM preprocessing step to transform irregular prose into a defined structure while validating the result afterward when source text is semistructured and deterministic parsing alone cannot reliably produce the required representation. The key requirement here is the text alone is usable but extra structured metadata improves downstream selection.
Answer C is incorrect because exact-content deduplication is designed to remove repeated identical text blocks before assembling the model context. That makes sense when duplicated passages waste tokens and can overweight one source. The question is centered on the text alone is usable but extra structured metadata improves downstream selection.
