AWS AI/ML Services: Curriculum Overview

This curriculum provides a structured path to understanding how Amazon Web Services (AWS) categorizes and delivers Artificial Intelligence (AI) and Machine Learning (ML) solutions. This is a critical component of the AWS Certified Cloud Practitioner (CLF-C02) exam, specifically falling under Domain 3: Cloud Technology and Services.

Prerequisites

Before diving into specific AI/ML services, learners should have a foundational understanding of the following:

• Cloud Fundamentals: Basic knowledge of the AWS Management Console and the concept of "Managed Services."
• Data Storage: Understanding of Amazon S3 as a data lake, as many AI services (like Textract and Kendra) pull data directly from S3 buckets.
• AI vs. ML Definitions:
  • Artificial Intelligence (AI): Systems designed to perform tasks that typically require human intelligence (reasoning, problem-solving).
  • Machine Learning (ML): A subset of AI involving algorithms that learn patterns from data to make predictions without explicit programming.

Module Breakdown

The curriculum is divided into three functional areas: Foundational Platforms, Specialized AI Services, and Search/Analysis.

Learning Objectives per Module

Module 1: Amazon SageMaker AI

• Objective: Understand the end-to-end ML lifecycle.
• Key Insight: SageMaker is for data scientists and developers to Build, Train, and Deploy models. It provides Jupyter Notebooks and no-code options like SageMaker Studio.

Module 2: Conversational & Language Services

• Objective: Differentiate between speech, text, and sentiment tools.
• Key Insight:
  • Lex: Build chatbots (same tech as Alexa).
  • Polly: Text-to-Speech (human-like voices).
  • Transcribe: Speech-to-Text (includes automatic redaction of sensitive info).
  • Comprehend: Natural Language Processing (NLP) to find sentiment and topics in text.

Module 3: Vision & OCR

• Objective: Identify objects and extract data from images/documents.
• Key Insight:
  • Rekognition: Identifies faces, objects, and inappropriate content in images/video.
  • Textract: Intelligently extracts text and tables from scanned documents, preserving relationships between data fields.

Success Metrics

To demonstrate mastery of this curriculum, you must be able to:

1. Map Business Needs to Services: Given a scenario (e.g., "We need to automate our call center routing"), identify the correct service (Amazon Lex).
2. Distinguish SageMaker from AI Services: Recognize that SageMaker is for custom model building, while the other services are pre-trained and available via API.
3. Identify Search Capabilities: Explain how Amazon Kendra provides natural language search across multiple data sources (S3, Salesforce, OneDrive).

[!IMPORTANT] For the exam, remember that Amazon Lex = Chatbots/Alexa technology, while Amazon Kendra = Intelligent/Enterprise Search.

Real-World Application

• Healthcare: Using Amazon Textract to digitize hand-written patient intake forms and Amazon Comprehend Medical to identify specific medical terms.
• Customer Support: Deploying an Amazon Lex bot to handle common FAQ questions, only escalating to a human agent when necessary.
• Security: Using Amazon Rekognition to verify identities via facial matching or to automatically flag prohibited content in user-uploaded videos.

Examples Section

Case Study: The "Smart" Office

To visualize how these services interact, consider an automated office environment:

1. The Receptionist (Lex + Polly): A visitor speaks to a kiosk. Lex understands the intent ("I'm here for a meeting"), and Polly responds in a natural voice to provide directions.
2. The Security Guard (Rekognition): Cameras analyze the visitor's face to check against a "known guest" list stored in S3.
3. The Research Assistant (Kendra): An employee asks the company portal, "What is our policy on remote work?" Kendra searches through thousands of PDFs and returns the specific paragraph containing the answer.
4. The Accountant (Textract): The employee scans a lunch receipt. Textract identifies the "Total" field and the "Tax" amount automatically, even if the receipt is wrinkled or poorly lit.

Checkpoint Questions

1. Which service should you use to convert a long audio recording of a meeting into a written transcript? (Answer: Amazon Transcribe)
2. If a company wants to build its own proprietary ML model using a no-code approach, which service is best? (Answer: Amazon SageMaker AI)
3. Which service can identify if a video contains unsafe or explicit content? (Answer: Amazon Rekognition)