CertBrain
LightNews LightNews
banner
certbrain.bsky.social
CertBrain
@certbrain.bsky.social
14 followers 79 following 23 posts
Get real spaced-repetition practice for the AWS AI Practitioner exam.
Posts Replies Media Videos
certbrain.bsky.social
CertBrain card of the day 📌

Conversational capabilities

Users ask questions in everyday language, and the application keeps track of earlier turns so follow-up questions make sense.

Not affiliated with AWS.

certbrain.app/decks/08493f...
Conversational capabilities: Users ask questions in everyday language, and the application keeps track of earlier turns so follow-up questions make sense.
November 1, 2026 at 10:00 PM
certbrain.bsky.social
Answer: A) Citations from RetrieveAndGenerate

Why, in the image.

Not affiliated with AWS.

certbrain.app/decks/c35f77...
Answer: A) Citations from RetrieveAndGenerate. RetrieveAndGenerate returns a generated answer together with citations that point to the source chunks, so staff can verify each answer. B is wrong because temperature affects wording consistency and adds no source information. C is wrong because longer answers don't tell users where facts came from. D is wrong because batch inference processes offline jobs and doesn't attach sources to answers.
November 1, 2026 at 8:10 PM
certbrain.bsky.social
CertBrain question of the day 🧠

A compliance team's assistant answers questions from internal policy documents. Staff must be able to check which document…

The choices are in the image. Answer later today.

Not affiliated with AWS.

certbrain.app/decks/c35f77...
A compliance team's assistant answers questions from internal policy documents. Staff must be able to check which document each answer came from. Which Amazon Bedrock Knowledge Bases capability supports this? A) Citations from RetrieveAndGenerate B) A lower temperature setting on the answering model C) A larger maximum output length for answers D) Batch inference over the policy documents
November 1, 2026 at 4:10 PM
certbrain.bsky.social
CertBrain card of the day 📌

Service role for Bedrock model customization

An IAM role Bedrock assumes to read your training data in S3 and write job results, so no access keys are shared.

Not affiliated with AWS.

certbrain.app/decks/751853...
Service role for Bedrock model customization: An IAM role Bedrock assumes to read your training data in S3 and write job results, so no access keys are shared.
October 31, 2026 at 9:10 PM
certbrain.bsky.social
Answer: D) Multi-turn conversation in plain language

Why, in the image.

Not affiliated with AWS.

certbrain.app/decks/08493f...
Answer: D) Multi-turn conversation in plain language. Understanding everyday language and keeping track of earlier turns is the conversational strength of LLMs, and Amazon Bedrock's Converse API is designed for multi-turn chat. A is wrong because image generation doesn't help answer account questions. B is wrong because generative AI isn't always cheaper than a simple rules engine; per-request costs can be higher. C is wrong because generative AI output can vary between runs, so identical answers aren't a strength it offers.
October 31, 2026 at 7:10 PM
certbrain.bsky.social
CertBrain question of the day 🧠

An online bank wants customers to ask account questions in everyday language and to ask follow-up questions that refer to…

The choices are in the image. Answer later today.

Not affiliated with AWS.

certbrain.app/decks/08493f...
An online bank wants customers to ask account questions in everyday language and to ask follow-up questions that refer to earlier answers, such as "what about my other card?". Today customers use a menu-driven help page. Which capability of generative AI MOST directly meets this need? A) Generating images from text descriptions B) Lower running cost than every rules engine C) Identical answers to identical questions D) Multi-turn conversation in plain language
October 31, 2026 at 3:10 PM
certbrain.bsky.social
Answer: D) Amazon Macie.

Macie finds personal, financial and credential data in S3. Inspector scans for software vulnerabilities, CloudTrail logs API calls, and Guardrails filter prompts and responses, not files.

Not affiliated with AWS.

certbrain.app/diagnostic
October 31, 2026 at 2:00 PM
certbrain.bsky.social
CertBrain card of the day 📌

Machine learning (ML)

The part of AI where a system learns patterns from example data instead of following rules a person wrote.

Not affiliated with AWS.

certbrain.app/decks/5d2196...
Machine learning (ML): The part of AI where a system learns patterns from example data instead of following rules a person wrote.
October 30, 2026 at 9:10 PM
certbrain.bsky.social
Answer: A) It drafts many variations quickly for people to review and test

Why, in the image.

Not affiliated with AWS.

certbrain.app/decks/08493f...
Answer: A) It drafts many variations quickly for people to review and test. The gain is speed and volume of drafts: generative AI makes forty variations practical, while people still approve them and testing still picks the winners. B is wrong because generated content can be inaccurate, so legal review stays. C is wrong because nothing guarantees the best copy first time; that's why the agency tests. D is wrong because a text generator doesn't measure how customers respond, so testing is still needed.
October 30, 2026 at 7:10 PM
certbrain.bsky.social
CertBrain question of the day 🧠

A marketing agency's writers produce three versions of ad copy per campaign. The agency wants to test about forty versions…

The choices are in the image. Answer later today.

Not affiliated with AWS.

certbrain.app/decks/08493f...
A marketing agency's writers produce three versions of ad copy per campaign. The agency wants to test about forty versions per campaign and keep the best performers, and its legal team must still approve every ad that runs. Which statement BEST describes the value generative AI adds here? A) It drafts many variations quickly for people to review and test B) It ends legal review because generated copy is always accurate C) It guarantees the highest-performing copy on the first attempt D) It replaces testing because the model predicts customer response
October 30, 2026 at 3:10 PM
certbrain.bsky.social
CertBrain practice question 🧠

Before fine-tuning, a retailer must find names, card numbers and credentials in thousands of S3 transcripts. Which service?

A) Amazon Inspector
B) AWS CloudTrail
C) Bedrock Guardrails
D) Amazon Macie

Answer tomorrow

Not affiliated with AWS.
October 30, 2026 at 2:00 PM
certbrain.bsky.social
CertBrain card of the day 📌

Amazon Macie

Scans data in Amazon S3 for sensitive information such as personal data, financial data and credentials, and flags risky bucket settings.

Not affiliated with AWS.

certbrain.app/decks/751853...
Amazon Macie: Scans data in Amazon S3 for sensitive information such as personal data, financial data and credentials, and flags risky bucket settings.
October 29, 2026 at 9:40 PM
certbrain.bsky.social
Answer: A) The model

Why, in the image.

Not affiliated with AWS.

certbrain.app/decks/5d2196...
Answer: A) The model. A model is what training produces: the learned patterns, saved so they can score new data during inference. B is wrong because the algorithm is the learning procedure the team ran, not its output; C is wrong because the training dataset is the two years of claims used to learn from; D is wrong because a feature list only names the input fields and cannot score anything on its own.
October 29, 2026 at 7:10 PM
certbrain.bsky.social
CertBrain question of the day 🧠

A data team ran a learning procedure over two years of past claims. The output is a saved file that the claims system now…

The choices are in the image. Answer later today.

Not affiliated with AWS.

certbrain.app/decks/5d2196...
A data team ran a learning procedure over two years of past claims. The output is a saved file that the claims system now calls to score each new claim for fraud risk. Which term describes that saved file? A) The model B) The algorithm C) The training dataset D) The feature list
October 29, 2026 at 3:10 PM
certbrain.bsky.social
New: a 4-week plan for the AWS AI Practitioner exam. Weighted by the 5 exam domains + a daily review habit to make it stick. Check out the guide below. Not affiliated with AWS.

certbrain.app/blog/a-four-...
October 29, 2026 at 2:30 PM
certbrain.bsky.social
One section at a time. Your place is saved as you go, and a new Aa button lets you dial in text size, spacing, width, and font. Read exactly how you like it. Not affiliated with AWS.

certbrain.app/blog/lessons...
October 29, 2026 at 2:30 PM
certbrain.bsky.social
Answer: B) Fairness.

Fairness is about how a system treats different groups. Controllability is steering the system, veracity is factual correctness, and transparency is telling people about it. None of those fixes the skew.

Not affiliated with AWS.

certbrain.app/diagnostic
October 29, 2026 at 2:00 PM
certbrain.bsky.social
Answer: A) Transparency

Why, in the image.

Not affiliated with AWS.

certbrain.app/decks/06c8d5...
Answer: A) Transparency. Transparency means giving people the information they need to decide how to engage with an AI system, such as disclosing it's an AI, stating its limits and offering an alternative. B is wrong because controllability is the team's ability to monitor and steer the system's behavior, not what customers are told. C is wrong because explainability is about understanding why a specific output was produced, which none of these steps provides. D is wrong because robustness is about correct outputs under unexpected inputs, which these disclosures don't change.
October 29, 2026 at 12:30 AM
certbrain.bsky.social
CertBrain question of the day 🧠

A bank tells customers when they are chatting with an AI assistant, publishes what the assistant can and can't help with…

The choices are in the image. Answer later today.

Not affiliated with AWS.

certbrain.app/decks/06c8d5...
A bank tells customers when they are chatting with an AI assistant, publishes what the assistant can and can't help with, and offers a clear way to reach a person instead. Which responsible AI dimension is the bank MAINLY addressing? A) Transparency B) Controllability C) Explainability D) Robustness
October 28, 2026 at 8:30 PM
certbrain.bsky.social
CertBrain is on Bluesky too: @certbrain.bsky.social. Same AIF-C01 practice questions and study tips as here. Not affiliated with AWS.
October 28, 2026 at 4:00 PM
certbrain.bsky.social
CertBrain practice question 🧠

A hiring tool recommends interviews for men far more often than for equally qualified women. Which responsible AI dimension is this MOST about?

A) Controllability
B) Fairness
C) Veracity
D) Transparency

Answer tomorrow

Not affiliated with AWS.
October 28, 2026 at 2:00 PM
certbrain.bsky.social
If you've unlocked CertBrain's full AIF-C01 course, you now have a code that gives a friend 30% off. When they unlock, you get 2 streak freezes. Not affiliated with AWS.

certbrain.app/blog/refer-a...
October 27, 2026 at 5:10 PM
certbrain.bsky.social
Answer: D) RAG with a Bedrock knowledge base.

Re-sync new rules into the knowledge base and the model answers from current text. Pre-training and fine-tuning are retraining; temperature changes wording, not knowledge.

Not affiliated with AWS.

certbrain.app/diagnostic
October 27, 2026 at 2:00 PM