Module 8: AI/ML and Data Analytics

Artificial Intelligence (AI)

A broad field focused on building intelligent computer systems capable of performing humanlike tasks.

Machine Learning (ML)

A type of AI that trains machines to perform tasks without explicit instructions by finding patterns in data to build ML models that make predictions or decisions.

Common ML Use Cases

Predict trends (stock prices), make decisions (route calls), detect anomalies (fraud detection).

AWS AI/ML Stack

Three tiers: AI services (pre-built models), ML services (Amazon SageMaker – build/train/deploy models), ML frameworks & infrastructure (custom models using AWS compute and frameworks).

Amazon Comprehend

Natural language processing (NLP) service that extracts key phrases, sentiment, and entities from documents. Use cases: content classification, customer sentiment analysis, compliance monitoring.

Amazon Polly

Converts text to lifelike speech with multiple languages and accents. Use cases: virtual assistants, e-learning, accessibility.

Amazon Transcribe

Converts speech to text with features like speaker identification and real-time transcription. Use cases: call transcription, subtitling, metadata generation.

Amazon Translate

Provides real-time and batch translation across multiple languages. Use cases: document translation, multi-language apps.

Amazon Kendra

Enterprise search service that uses NLP to deliver context-aware answers instead of keyword matches. Use cases: chatbots, intelligent search.

Amazon Rekognition

Analyzes images and videos to detect objects, people, text, scenes, activities. Use cases: content moderation, identity verification, media analysis.

Amazon Textract

Detects and extracts typed and handwritten text from documents, forms, tables. Use cases: form processing in finance, healthcare, government.

Amazon Lex

Adds voice and text conversational interfaces using NLU + ASR. Use cases: virtual assistants, FAQ bots, chatbots.

Amazon Personalize

Uses historical data to generate personalized recommendations. Use cases: product recommendations, personalized streaming.

Amazon SageMaker

Fully managed service to build, train, deploy ML models. Provides IDE, experiment tracking, pre-trained models, and workflow monitoring.

ML Frameworks

Software libraries like PyTorch, MXNet, TensorFlow for building custom ML models.

AWS ML Infrastructure

ML-optimized EC2 instances, Amazon EMR, Amazon ECS for high-performance, flexible ML workloads.

Deep Learning (DL)

A subset of ML that uses artificial neural networks with layered neurons to produce complex models.

Generative AI

Deep learning models called foundation models (FMs) trained on massive datasets. Can handle multiple tasks, including text, image, video, and music generation.

Large Language Models (LLMs)

A type of foundation model trained to understand and generate human language.

Amazon SageMaker JumpStart

Hub within SageMaker with pre-built ML solutions across domains (vision, NLP, tabular data). Accelerates model building & deployment.

Amazon Bedrock

Fully managed service for building generative AI apps using foundation models from Amazon & partners (Claude, Stable Diffusion). Supports experimentation, fine-tuning, and integration.

Amazon Q

Interactive AI assistant for organizations.

Amazon Q Business

Answers questions and takes actions using company data. Use cases: workflow automation, data insights.

Amazon Q Developer

Provides code recommendations, generates functions and logic blocks, integrates with IDEs. Use cases: faster code generation, automated reviews.

ETL Process

Extract, Transform, Load data into a warehouse or platform for analysis. Data pipelines automate ETL.

Data Analytics

Transforms raw data into insights. Use cases: loan decisions, clinical trial analysis, risk modeling.

Amazon Kinesis Data Streams

Real-time data ingestion service for streams/sensors. Serverless, auto-scaling.

Amazon Data Firehose

Near real-time data ingestion service that delivers data to lakes, warehouses, analytics tools within seconds.

Amazon S3 for Data Lakes

Object storage that can house virtually unlimited structured/unstructured data, scales automatically.

Amazon Redshift

Fully managed data warehouse storing petabytes of data. Ideal for large-scale analytics with columnar storage and parallel processing.

AWS Glue Data Catalog

Provides centralized metadata repository for data discovery, improves data pipeline efficiency.

AWS Glue

Fully managed ETL service for data cleaning and transformation using the Glue Data Catalog.

Amazon Elastic MapReduce (EMR)

Manages big data frameworks (Spark, Hadoop, Hive) for large-scale data processing. Handles provisioning, cluster management, scaling.

Amazon Athena

Serverless SQL query service to analyze data stored in S3 or other sources. Pay only for queries run.

Amazon QuickSight

Tool for creating interactive dashboards & reports. Supports natural language queries (Amazon Q in QuickSight).

Amazon OpenSearch Service

Provides real-time search & visualization for logs, metrics, traces. Supports natural language queries and unified dashboards.