aws_ai_practitioner.md

AWS AI Practitioner

1. Introduction to Artificial Intelligence (AI)

   • Definition:
     • AI refers to the simulation of human intelligence in machines designed to think and act like humans.
     • Capabilities include understanding natural language, recognizing patterns, solving problems, and making decisions.
   • Examples of AI Applications:
     • Personal Assistants:
       • Virtual assistants like Alexa and Siri that perform tasks such as setting reminders, answering questions, and controlling smart home devices.
     • Fraud Detection:
       • Systems designed to identify and prevent fraudulent activities by analyzing transaction data in real-time to detect anomalies.
     • Medical Imaging:
       • AI applications that analyze medical images like X-rays, MRIs, and CT scans to assist in diagnosis and treatment planning.
     • Manufacturing:
       • Uses AI for quality control by identifying defects in products and for predictive maintenance to anticipate equipment failures.
     • Customer Support:
       • Automated chatbots that handle customer queries and provide product recommendations.
     • Predictive Analytics:
       • Utilizing historical data to forecast future trends and demands, aiding in strategic planning and decision-making.
   • Key Concepts:
     • Machine Learning (ML): A subset of AI that involves algorithms learning from data to make decisions without explicit programming.
     • Deep Learning: A further subset of ML that uses neural networks with many layers (deep networks) to analyze complex data patterns.
     • Generative AI: A branch of AI that focuses on creating new content, such as text, images, or code, by learning from existing data, often using models like neural networks.

2. Machine Learning (ML)

   • Definition:
     • ML is a method of data analysis that automates the building of analytical models based on the idea that systems can learn from data, identify patterns, and make decisions with minimal human intervention.
   • Types of Data:
     • Structured Data:
       • Data that is organized in a defined manner (e.g., databases, spreadsheets).
       • Examples: Sales data, customer information.
     • Semi-Structured Data:
       • Partially organized data that doesn't fit into a relational database but has some organizational properties.
       • Examples: JSON files, XML documents.
     • Unstructured Data:
       • Data that does not have a predefined structure.
       • Examples: Text data (emails, social media posts), images, videos.
   • Training Process:
     • Involves feeding large amounts of data to an algorithm so that it can learn to make predictions or decisions.
     • Algorithms:
       • Mathematical models that process the data to find patterns or relationships.
     • Features:
       • Measurable properties or characteristics of the data used as input for algorithms.
     • Inference:
       • The process of using the trained model to make predictions on new, unseen data.
   • Machine Learning Styles:
     • Supervised Learning:
       • Trains on labeled data where the output is known.
       • Examples:
         • Image Classification: Identifying objects within images.
         • Spam Detection: Classifying emails as spam or not spam.
     • Unsupervised Learning:
       • Trains on unlabeled data to find hidden patterns.
       • Examples:
         • Clustering Analysis: Grouping data points based on similarity (e.g., customer segmentation).
         • Anomaly Detection: Identifying unusual data points (e.g., fraud detection).
     • Reinforcement Learning:
       • Trains an agent to make decisions through trial and error by receiving rewards or penalties.
       • Examples:
         • Game Playing: Training AI to play games like chess or Go.
         • Robotics: Teaching robots to navigate environments.

3. Deep Learning

   • Definition:
     • A subset of machine learning that utilizes neural networks with multiple layers (hence "deep") to model complex patterns in large datasets.
   • Neural Network Structure:
     • Input Layer:
       • Receives the initial data (e.g., pixels in an image, words in a sentence).
     • Hidden Layers:
       • Multiple layers where the data is processed. Each layer extracts features and passes them to the next layer.
       • Types of Layers:
         • Dense (Fully Connected) Layers: Every neuron is connected to every neuron in the next layer.
         • Convolutional Layers: Used primarily in image processing to detect spatial hierarchies.
         • Recurrent Layers: Used in sequence data to remember previous inputs (e.g., LSTMs for text).
     • Output Layer:
       • Produces the final prediction or classification (e.g., classifying an image as a dog or cat).
   • Applications:
     • Image Classification:
       • Identifying objects within images, such as recognizing different species of animals in photos.
     • Natural Language Processing (NLP):
       • Understanding and generating human language, such as translating languages or summarizing text.
   • Deep Learning vs. Traditional ML:
     • Data Type:
       • Traditional ML: Structured and labeled data.
       • Deep Learning: Unstructured data like images, text, and audio.
     • Feature Extraction:
       • Traditional ML: Requires manual feature selection and extraction.
       • Deep Learning: Automatically extracts features from raw data.
     • Computation Cost:
       • Traditional ML: Generally lower computational cost.
       • Deep Learning: Higher computational cost due to large datasets and complex models.
     • Use Cases:
       • Traditional ML: Predictive analytics, classification, recommendation.
       • Deep Learning: Image recognition, speech recognition, language translation.

4. Generative AI

   • Definition:
     • Refers to models that generate new content based on training data.
   • Techniques:
     • Transformers:
       • A type of model architecture that processes sequences of data (e.g., sentences) in parallel, making them efficient for training on large datasets.
       • Components of Transformers:
         • Self-Attention Mechanism: Weighs the importance of different parts of the input when generating output.
         • Encoder-Decoder Architecture: Consists of encoder layers to process input and decoder layers to generate output.
         • Positional Encoding: Encodes the relative position of each token in a sequence to preserve order.
   • Applications:
     • Content Creation:
       • Writing articles, generating images, composing music.
     • Language Models:
       • Understanding and generating human language, such as in chatbots and translation services.
   • Core Components:
     • Models:
       • Built using neural networks, trained to generate output resembling the input data.
     • Tokenization:
       • Converts human text into vectors called token IDs, which represent words or units in the model's vocabulary.
     • Embeddings:
       • Numerical vector representations of tokens, capturing semantic meaning and context.
     • Self-Attention Mechanism:
       • Computes query, key, and value vectors for each token to determine attention weights.
     • Positional Encoding:
       • Encodes the relative position of each token to maintain the structure and order of sentences.
   • In-Context Learning:
     • Few-Shot Learning:
       • Provides a few examples within a prompt to guide the model in generating better outputs.
       • Example: Showing the model a few translated sentences to improve its translation capability.
     • Zero-Shot Learning:
       • The model performs a task it hasn't been explicitly trained for, without examples.
       • Example: Asking a model to generate a summary without providing any prior examples.
     • One-Shot Learning:
       • Provides only one example to learn from.
       • Example: Teaching the model to classify a rare object with a single labeled example.

5. Guidelines for Responsible AI

   • Development of Responsible AI Systems:
     • Ensuring AI systems are ethical, transparent, and fair.
     • Principles:
       • Fairness: AI should be unbiased and treat all individuals equally.
       • Transparency: The workings of AI models should be understandable.
       • Robustness: AI systems should be resilient and handle unexpected situations gracefully.
       • Privacy and Security: Protecting user data and ensuring compliance with privacy regulations.
   • Transparent and Explainable Models:
     • Importance of creating AI models that are interpretable and explainable.
     • Techniques for Explainability:
       • LIME (Local Interpretable Model-agnostic Explanations): Provides local explanations for individual predictions.
       • SHAP (SHapley Additive exPlanations): Calculates the contribution of each feature to the model's prediction.
       • Integrated Gradients: Attributes the prediction of a model to its input features by computing gradients.

6. Security, Compliance, and Governance for AI Solutions

   • Methods to Secure AI Systems:
     • Shared Responsibility Model:
       • AWS Responsibilities: Infrastructure security, service management.
       • Customer Responsibilities: Service configuration, application security.
     • Identity and Access Management (IAM):
       • IAM Users: Represent individuals needing access to AWS services.
       • IAM Groups: Collections of users with similar permissions.
       • IAM Roles: Temporary access permissions for AWS resources.
       • Principle of Least Privilege: Grant minimal permissions necessary.
     • Data Encryption:
       • Data at Rest: Encryption of stored data.
       • Data in Transit: Encryption during data transfer.
       • AWS Key Management Service (KMS): Management of encryption keys.
     • Logging and Monitoring:
       • AWS CloudTrail: Captures and logs API calls.
       • Amazon SageMaker Role Manager: Simplifies the creation of IAM roles for ML tasks.
   • Governance and Compliance Regulations for AI Systems
     • AWS Compliance Tools:
       • AWS Audit Manager: Automates compliance audits and evidence collection.
       • AWS Config: Monitors resource configurations and evaluates compliance.
       • Amazon Inspector: Provides automated security assessments.
       • AWS Trusted Advisor: Offers guidance on security best practices.

7. Types of Machine Learning Problems

   • Supervised Learning:
     • Definition: The model is trained on a labeled dataset, where each training example is paired with an output label.
     • Types of Supervised Learning:
       • Classification:
         • Binary Classification: Categorizes data into two classes (e.g., spam vs. not spam emails).
         • Multiclass Classification: Categorizes data into more than two classes (e.g., categorizing news articles into sports, finance).
       • Regression:
         • Linear Regression: Predicts a continuous output with a linear relationship between input and output.
         • Multiple Linear Regression: Uses multiple input variables to predict the output.
         • Logistic Regression: Used for binary classification tasks, predicting the probability of an event occurring.
   • Unsupervised Learning:
     • Definition: The model is given data without explicit instructions on what to do with it, identifying underlying patterns or structures.
     • Clustering:
       • K-Means Clustering: Divides data into a predefined number of clusters based on similarity.
       • Hierarchical Clustering: Builds a tree of clusters based on data similarity.
     • Anomaly Detection: Identifies rare items or events that do not conform to expected patterns (e.g., fraud detection).
   • Semi-Supervised Learning:
     • Definition: A blend of supervised and unsupervised learning, where the model is trained on a small amount of labeled data and a larger amount of unlabeled data.
   • Reinforcement Learning:
     • Definition: An agent learns to make decisions by performing actions in an environment to maximize cumulative rewards.
     • Examples: Game playing, robotics.

8. Model Deployment

   • Batch vs. Real-Time Inference:
     • Batch Inference:
       • Ideal for large numbers of inferences where results can be delayed (e.g., overnight processing).
       • Cost-effective as resources are used intermittently.
     • Real-Time Inference:
       • Suitable for immediate responses to client requests, often via a REST API.
       • Deployed models respond immediately, ideal for applications like chatbots.
   • Deployment Options:
     • AWS API Gateway & Lambda:
       • API Gateway: Handles client interactions and passes requests to Lambda running the model.
     • Docker Containers:
       • Used for deploying models, offering versatility across AWS services (ECS, EKS, Lambda, EC2).
     • Amazon SageMaker:
       • Provides managed endpoints for various inference types (batch, asynchronous, serverless, real-time).
       • Simplifies deployment by managing infrastructure, scalability, and updates.

9. Model Monitoring

   • Performance Degradation:
     • Over time, model performance may degrade due to factors like data quality, model quality, and bias.
     • Mitigation Strategies:
       • Retraining models with new data, adjusting algorithms, or updating features.
   • Monitoring Systems:
     • Data & Concept Drift:
       • Detects significant changes in data distribution (data drift) and changes in target variable properties (concept drift).
     • Amazon SageMaker Model Monitor:
       • Monitors models in production, detects errors, and compares data against a baseline.
       • Sends alerts via CloudWatch, potentially triggering re-training cycles.
   • Automation & MLOps:
     • MLOps:
       • Incorporates DevOps practices into ML model development, focusing on automating tasks, ensuring version control, and monitoring deployments.
       • Improves productivity, repeatability, reliability, compliance, and data quality.
     • Amazon SageMaker Pipelines:
       • Facilitates the orchestration of ML pipelines, enabling the deployment of models and tracking lineage.

10. Model Evaluation Metrics

    • Classification Metrics:
      • Confusion Matrix:
        • True Positive (TP): Correctly predicted positive cases.
        • True Negative (TN): Correctly predicted negative cases.
        • False Positive (FP): Incorrectly predicted positive cases.
        • False Negative (FN): Incorrectly predicted negative cases.
      • Accuracy: Measures the percentage of correct predictions. Suitable for balanced datasets.
      • Precision: Focuses on the accuracy of positive predictions. Important when minimizing false positives.
      • Recall: Measures the ability to detect all actual positives. Used when minimizing false negatives is critical.
      • F1 Score: Balances precision and recall. Ideal when both metrics are important.
      • AUC-ROC: Evaluates binary classification models by plotting true positive rate against false positive rate across thresholds.
    • Regression Metrics:
      • Mean Squared Error (MSE): Average of squared differences between predictions and actual values. Sensitive to outliers.
      • Root Mean Squared Error (RMSE): Square root of MSE, easier to interpret as it's in the same units as the dependent variable.
      • Mean Absolute Error (MAE): Average of absolute errors, less sensitive to outliers than MSE.
    • Business Metrics:
      • Return on Investment (ROI): Measures the profitability of an investment.
      • Cost Reduction: Quantifies the savings achieved through AI solutions.
      • Increased Sales: Evaluates the impact of AI solutions on revenue growth.
      • AWS Cost Explorer with Cost Allocation Tags: Monitors project expenses.
    • Generative AI Metrics:
      • ROUGE (Recall-Oriented Understudy for Gisting Evaluation): Measures the quality of summarization and translation by comparing generated text to reference text.
      • BLEU (Bilingual Evaluation Understudy): Evaluates machine translation by comparing the model's translations to human translations.
      • GLUE (General Language Understanding Evaluation): A benchmark that tests various language understanding tasks like sentiment analysis.
      • SuperGlue: Extends GLUE by adding tasks that require complex reasoning and understanding, like reading comprehension.
      • MMLU (Massive Multitask Language Understanding): Tests a model's knowledge and problem-solving skills across diverse topics, from history to mathematics.
      • BIG-bench: Challenges models with tasks beyond current capabilities, such as advanced reasoning and specialized knowledge.
      • HELM (Holistic Evaluation of Language Models): Focuses on improving model transparency and evaluates performance on tasks like summarization and sentiment analysis.

11. AWS AI Services

    • Computer Vision Services:
      • Amazon Rekognition:
        • Deep learning service for computer vision tasks.
        • Use Cases: Face recognition, object detection, content moderation, real-time video analysis.
      • Amazon Textract:
        • Extracts text, handwriting, forms, and tables from scanned documents.
        • Use Cases: Automating document processing (e.g., invoices, forms).
    • Natural Language Processing (NLP) Services:
      • Amazon Comprehend:
        • NLP service that discovers insights and relationships in text.
        • Use Cases: Sentiment analysis, PII detection, entity recognition.
      • Amazon Lex:
        • Builds voice and text interfaces using Amazon Alexa technology.
        • Use Cases: Chatbots, interactive voice response systems for customer service.
      • Amazon Polly:
        • Converts text into natural-sounding speech in multiple languages.
        • Use Cases: Text-to-speech conversion for audio content, enhancing accessibility and engagement.
      • Amazon Kendra:
        • ML-powered search service for enterprise systems.
        • Use Cases: Intelligent search with natural language queries.
      • Amazon Transcribe:
        • Converts spoken language into text (speech-to-text).
        • Use Cases: Real-time transcription, captioning for live or recorded audio/video.
    • Personalization & Recommendation Services:
      • Amazon Personalize:
        • Provides personalized recommendations for customers.
        • Use Cases: Product/content recommendations, targeted marketing campaigns.
    • Translation Services:
      • Amazon Translate:
        • Neural machine translation for text across 75 languages.
        • Use Cases: Real-time translation in chat applications, multilingual content creation.
    • Forecasting & Planning Services:
      • Amazon Forecast:
        • AI service for time series forecasting.
        • Use Cases: Demand forecasting, inventory management, financial planning.
    • Fraud Detection Services:
      • Amazon Fraud Detector:
        • Detects potentially fraudulent online activities using pre-trained models.
        • Use Cases: Preventing online payment fraud, detecting fake accounts, account takeover prevention.
    • Generative AI Services:
      • Amazon Bedrock:
        • Service to build generative AI applications using foundation models from top AI providers.
        • Use Cases: Content creation, image generation, retrieval augmented generation (RAG) for enhanced model accuracy.
    • Custom ML Development:
      • Amazon SageMaker:
        • Comprehensive service for building, training, and deploying custom ML models.
        • Use Cases: Custom model development for predictive analytics, large-scale data processing, real-time inference.

12. Amazon SageMaker Services

    • SageMaker Ground Truth:
      • A data labeling service to build highly accurate training datasets for machine learning quickly.
      • Features: Human-in-the-loop labeling, integration with other AWS services, automated data labeling using machine learning models.
    • SageMaker Canvas:
      • Enables business analysts to build machine learning models and generate accurate predictions without writing code.
      • Features: No-code interface, automated model generation, supports structured data.
    • SageMaker Experiments:
      • A tool to organize, track, compare, and evaluate machine learning experiments.
      • Features: Experiment tracking, lineage tracking, comparison of experiment results, integration with SageMaker Studio.
    • SageMaker Model Monitor:
      • Monitors deployed models in production for data and model quality issues and automatically detects and alerts on potential problems.
      • Features: Real-time monitoring, alerting via CloudWatch, integration with SageMaker Studio for visualization, supports custom rules and built-in monitors.
    • SageMaker Pipelines:
      • A service to build, automate, and manage end-to-end machine learning workflows.
      • Features: Workflow orchestration, model deployment, lineage tracking, integration with SageMaker Studio, Python SDK, JSON-based pipeline definition, supports conditional logic.
    • SageMaker Model Registry:
      • A centralized repository to store, version, and manage machine learning models.
      • Features: Model versioning, model lineage tracking, integration with deployment pipelines, support for multiple model versions.
    • SageMaker Feature Store:
      • A purpose-built repository for storing, retrieving, and sharing machine learning features.
      • Features: Feature definition storage, real-time and offline retrieval, integration with SageMaker Pipelines and SageMaker Studio, versioning of feature definitions.
    • SageMaker Inference Recommender:
      • Helps select the best compute instance and configuration for inference workloads by running benchmark tests on different configurations.
      • Features: Instance type recommendation, configuration testing, support for different inference options, integration with SageMaker deployment.
    • SageMaker Serverless Inference:
      • Allows serving machine learning models without managing infrastructure, automatically scaling based on traffic patterns.
      • Features: No need for provisioning instances, automatic scaling, cost-effective for intermittent workloads, leverages AWS Lambda.
    • SageMaker Real-Time Inference:
      • Provides persistent endpoints for real-time inference that are fully managed and can automatically scale.
      • Features: Low-latency real-time responses, persistent endpoints, support for auto-scaling, integration with other AWS services like API Gateway.
    • SageMaker Batch Transform:
      • A service for offline inference that processes large datasets in batches.
      • Features: Suitable for large datasets, supports gigabyte-scale data, no need for persistent endpoints, integration with S3 for input/output data.
    • SageMaker Asynchronous Inference:
      • Supports workloads that involve large payloads or have long inference processing times, decoupling request and response so clients don't have to wait for the inference response.
      • Features: Asynchronous response handling, decoupling request and response, support for large payloads, storage of results in S3, cost-effective for long-running or large-payload inferences.

13. Foundational Models

    • Selection Criteria for Pre-trained Models:
      • Cost: Consider the expense of training the model, including hardware, storage, and computational resources.
      • Latency Constraints: For real-time applications, the model must provide rapid responses.
      • Modalities Supported: Models may handle different types of data (text, image, etc.) and may require ensemble methods to improve performance.
      • Architecture and Complexity: More complex models may offer higher accuracy but require more computational resources.
      • Performance Metrics: Evaluate models using metrics like accuracy, precision, recall, F1 score, RMSE, MAP, MAE.
    • Biases in Training Data:
      • Bias Mitigation: Address biases present in training data to ensure ethical and fair outcomes.
      • Ethical Considerations: Make informed decisions about model selection and fine-tuning with a focus on minimizing biases.
    • Availability and Compatibility:
      • Model Repositories: Check if the model is available on platforms like TensorFlow Hub, PyTorch Hub, Hugging Face.
      • Compatibility: Ensure the model aligns with your framework, language, and environment.
    • Customization and Explainability:
      • Customization Techniques:
        • Model Fine-Tuning: Adjusting a pre-trained model on new data to improve task-specific performance.
        • Transfer Learning: Adapting a pre-trained model to a new but related task.
        • Meta Learning: Models learn to adapt to new tasks quickly.
        • Self-Supervised Learning: Models learn to predict parts of their input data, creating labeled data from raw data.
      • Explainability Tools:
        • LIME, SHAP, Integrated Gradients: Techniques for interpreting model predictions.
    • Inference Parameters:
      • Temperature: Controls the randomness of responses. Higher values increase diversity, lower values make the output more focused and deterministic.
      • Top K: Limits the number of top predictions considered during generation, reducing randomness.
      • Top P (Nucleus Sampling): Uses cumulative probability to determine the response space, dynamically choosing the set of likely next words.
      • Response Length: Sets limits on the length of model outputs to prevent overly long or short responses.
      • Penalties: Adjusts the model's tendency to repeat the same output (repetition penalty) or to continue a thought (presence penalty).
    • Evaluation Metrics for Generative AI:
      • ROUGE: Evaluates the quality of text summarization.
      • BLEU: Measures the accuracy of machine translation.
      • GLUE: Benchmarks for general language understanding.
      • SuperGlue: Extends GLUE with more challenging language understanding tasks.
      • MMLU: Tests broad knowledge and problem-solving skills.
      • BIG-bench: Evaluates models on tasks that are beyond current capabilities.
      • HELM: Focuses on transparency and bias detection in AI outputs.

14. Prompt Engineering Techniques

    • Introduction to Prompts:
      • Definition and components of a prompt.
    • Prompting Techniques:
      • Few-Shot Prompting: Providing a few examples to guide the model.
        • Example: Translate the following sentences into French.
      • Zero-Shot Prompting: Asking the model to perform a task without examples.
        • Example: Translate "good morning" to Spanish.
      • One-Shot Prompting: Providing a single example.
        • Example: Show how to solve a single math problem to guide the model.
      • Chain-of-Thought Prompting: Breaking down complex tasks into intermediate steps to improve coherence.
        • Example: Provide step-by-step reasoning for a scientific explanation.
      • Prompt Tuning: Using continuous embeddings optimized during training to improve model outputs.
    • Best Practices:
      • Be Specific: Define clear instructions and examples.
      • Include Examples: Guide the model with sample inputs and outputs.
      • Experiment and Iterate: Test and refine prompts to enhance model performance.
      • Use Multiple Comments: Provide context without cluttering the prompt.
      • Add Guardrails: Implement safety measures to manage AI interactions.
    • Risks and Limitations:
      • Prompt Injection: Manipulating prompts to produce unintended outputs.
      • Jailbreaking: Bypassing safety mechanisms set by prompt engineers.
      • Hijacking: Changing the original prompt with new instructions.
      • Poisoning: Embedding harmful instructions in various inputs.

15. Vector Databases and Retrieval Augmented Generation (RAG)

    • Vector Databases:
      • Function: Store data as numerical vectors for efficient lookups and enhance model capabilities by providing relevant data.
      • AWS Services for Vector Search:
        • Amazon OpenSearch Service, Amazon Aurora, Redis, Amazon Neptune, Amazon DocumentDB, Amazon RDS with PostgreSQL.
    • Retrieval Augmented Generation (RAG):
      • Components:
        • Retriever: Searches knowledge base for relevant data.
        • Generator: Produces outputs based on the retrieved data.
      • Applications:
        • Question Answering: Enhances model responses by integrating external knowledge.
        • Content Generation: Uses external data to improve content accuracy.

16. Overview of Responsible AI

    • Core Dimensions:
      • Fairness: Ensures equitable treatment across diverse groups.
      • Explainability: Provides clear reasons for AI decisions.
      • Robustness: Ensures tolerance to failures and minimizes errors.
      • Privacy: Protects user data and ensures PII is not exposed.
      • Governance: Meets compliance and risk management standards.
      • Transparency: Clearly communicates model capabilities and risks.

17. Methods to Secure AI Systems

    • Shared Responsibility Model:
      • AWS Responsibilities: Security of the cloud infrastructure.
      • Customer Responsibilities: Security within the cloud.
    • IAM (Identity and Access Management):
      • Purpose: Manages access to AWS resources, including user creation, permissions, and MFA.
      • Root User: Initial account with unrestricted access; best practices include minimizing usage and enabling MFA.
      • IAM Users and Groups: Best practices for managing user access.
      • IAM Roles: Reducing risk by providing temporary access.
    • Data Encryption:
      • Types: Data at Rest and Data in Transit.
      • AWS KMS (Key Management Service): Manage and control encryption keys.
    • S3 Block Public Access: Prevents public access to S3 buckets and objects.
    • SageMaker Role Manager: Simplifies role creation for SageMaker tasks.

18. Compliance Tools and AWS Services

    • AWS Audit Manager: Maps compliance requirements to AWS usage data and produces assessment reports.
    • AWS Config: Monitors resource configurations and compliance.
    • Amazon Inspector: Assesses security vulnerabilities in applications and containers.
    • AWS Trusted Advisor: Provides recommendations for cost optimization, performance, security, and operational excellence.
    • AWS Glue DataBrew: Visual data preparation and quality management tools.
    • AWS Glue Data Quality: Sets data quality rules and detects anomalies.