AWS CERTIFIED AI PRACTITIONER
UPDATED SCRIPT 2026 PRACTICE
SOLUTIONS GRADED A+
● Amazon Bedrock - Base Foundation Model (FM). Answer: A
selection of industry-leading models that can be used to perform tasks
like image generation, text summarization, etc. used to generate data, we
rely on this
trained on a wide variety of input data, costs tens of millions of $ to
train.
● Large Language Models (LLM). Answer: type of AI designed to
generate coherent human-like text (GPT4 for ChatGBT/Open AI)
trained on large corpus of text data, usually very big models (billions of
parameters)
can perform language-related tasks (translation, summarization, question
answering)
Interact with this by giving a prompt
, It is *Non-deterministic* which means the generated text may be
different for every user that uses the same prompt
● Amazon Bedrock. Answer: A service that builds Gen AI applications
on AWS. No servers for you to manage (fully-managed)
keep control of your data used to train the model. Pay-per-use model
Makes a copy of the Foundation Model (FM) only to you, which you
can further fine-tube with your own data
Access to wide range of FM (Meta, AmaZon, cohere, etc.)
*None of your data is used to train the FM*
● Amazon Titan. Answer: This refers to AWS large language models
(LLMs) as part of their *Bedrock* suite of gen AI services.
High-performing Foundation Models directly from AWS. Can do
images, texts, and also have multi-model choices
Is a Virtual Private Cloud (VPC) and can be customized with your own
data, provides control and customization of infrastructure. (compared to
Bedrock which is serverless & fully managed serviced)
UPDATED SCRIPT 2026 PRACTICE
SOLUTIONS GRADED A+
● Amazon Bedrock - Base Foundation Model (FM). Answer: A
selection of industry-leading models that can be used to perform tasks
like image generation, text summarization, etc. used to generate data, we
rely on this
trained on a wide variety of input data, costs tens of millions of $ to
train.
● Large Language Models (LLM). Answer: type of AI designed to
generate coherent human-like text (GPT4 for ChatGBT/Open AI)
trained on large corpus of text data, usually very big models (billions of
parameters)
can perform language-related tasks (translation, summarization, question
answering)
Interact with this by giving a prompt
, It is *Non-deterministic* which means the generated text may be
different for every user that uses the same prompt
● Amazon Bedrock. Answer: A service that builds Gen AI applications
on AWS. No servers for you to manage (fully-managed)
keep control of your data used to train the model. Pay-per-use model
Makes a copy of the Foundation Model (FM) only to you, which you
can further fine-tube with your own data
Access to wide range of FM (Meta, AmaZon, cohere, etc.)
*None of your data is used to train the FM*
● Amazon Titan. Answer: This refers to AWS large language models
(LLMs) as part of their *Bedrock* suite of gen AI services.
High-performing Foundation Models directly from AWS. Can do
images, texts, and also have multi-model choices
Is a Virtual Private Cloud (VPC) and can be customized with your own
data, provides control and customization of infrastructure. (compared to
Bedrock which is serverless & fully managed serviced)
