[NEW] AWS Certified AI Practitioner AIF-C01

Explore the fundamentals of machine learning and artificial intelligence, including generative AI and traditional ML. Discover text, image, audio, and code generation, responsible use, and AWS services in AI.

Explore core AI terms and their relationships, from AI as an umbrella to machine learning, deep learning, neural networks, generative AI, computer vision, and natural language processing.

Explore how AI models and algorithms power computer vision and natural language processing, covering training and inference, bias, fairness, model fit, and large language models.

Differentiate artificial intelligence, machine learning, and deep learning, and explain how data, complexity, and data-driven learning shape their applications from virtual assistants to computer vision and natural language processing.

Explore AI inferences and training data quality, emphasize garbage in, garbage out, and distinguish labeled vs unlabeled data, supervised vs unsupervised vs reinforcement learning, plus structured, semi-structured, unstructured data.

Explore supervised, unsupervised, and reinforcement learning and how models learn from labeled or unlabeled data. See examples like image classification, clustering, and anomaly detection.

Explore how a trained model performs inference, comparing batch, real-time, and edge inference, with examples like historical sales data and drones performing real-time decisions.

Evaluate when AI and ML are not appropriate by weighing costs, data quality, and expected outcomes, and consider traditional or rule-based solutions for projects with limited data or budgets.

Explore practical AI use cases and AWS managed ML services, including Rekognition, Textract, Comprehend, Lex, Transcribe, and Polly, to leverage pre-trained capabilities and avoid custom model training.

Discover AWS managed AI/ML services across customer experience and business metrics, featuring Kendra, Personalize, Translate, Forecast, Fraud Detector, Bedrock, Queue, Titan image generator, and SageMaker.

Explore the machine learning development life cycle, from problem definition and data preparation to model training, implementation, and monitoring, with an emphasis on iterative refinement toward clear business objectives.

Learn how to collect, clean, and prepare training data for a movie recommendation system using ETL pipelines, data labeling, feature engineering, and train/validation/test splits with AWS Glue and SageMaker.

Train, tune, and evaluate a movie recommender model with iterative parameter optimization and parallel experiments using Amazon SageMaker, S3 data, and Docker containers.

Evaluate model performance using false positive rate, true negative rate, AUC/ROC across thresholds, and regression with MSE and R-squared, including business ROI with AWS Redshift and QuickSight.

Deploy the trained movie recommendation model by choosing batch or real-time inference, and expose it via a Rest API using AWS services such as SageMaker, Lambda, and API gateway.

Master the fundamentals of ml operations (MLOps), including version control, automation, CI/CD, and governance, to manage the ml lifecycle from development to deployment and monitoring.

Explore foundational models as the cornerstone of generative AI, pre-trained on internet-scale data to power text generation, image creation, and multimodal tasks, with Bert, GPT three, and Dall-E as examples.

Compare unimodal and multimodal models in generative ai, covering text, images, and audio. Explore applications like image captioning, visual question answering, and text-to-image synthesis with Dall-E, Stable diffusion, and Midjourney.

Learn how generative adversarial networks (GANs) use a generator and a discriminator in a competitive training loop to produce realistic data, including images, from random noise.

Explore variations of generative adversarial networks, including progressive GANs, image-to-image translation, cycle GANs, and text-to-image synthesis. Discover applications in super resolution, face painting task, and speech synthesis from text.

Explore diffusion models as powerful generative tools that learn to reverse noise to create high-quality images from noise or text, using stable diffusion and Amazon SageMaker Jumpstart.

Explore how generative ai models create images, videos, music, and text. Identify AWS services like Amazon Polly, Lex, Translate, Comprehend, Textract, CodeWhisperer, Kendra, and Personalize that enable these use cases.

Identify the business value of generative AI by defining metrics like multi-domain performance, efficiency, conversion rate, average revenue per user, customer lifetime value, and accuracy across marketing, service, and sales.

Explore Amazon SageMaker Jumpstart with pre-trained models and templates, Bedrock foundational models, Padirac playground, and Amazon Queue Business analytics for generative ai applications.

Explore the advantages of AWS AI solutions, including Amazon Q developer, SageMaker, and Bedrock, for fast development, scalable deployment, and access to pre-trained models and APIs.

Explore the cost tradeoffs of AWS generative AI services, balancing responsiveness, redundancy, regional coverage, throughput, pricing models, and custom models to optimize performance and costs.

Discover AWS AI/ML/Gen AI services from SageMaker to Comprehend, Translate, Textract, Lex, Polly, Transcribe, Rekognition, Kendra, Personalize, DeepRacer, Bedrock, Jumpstart, and Q for building advanced AI applications.

Choose and customize pre-trained models for problems using retrieval augmented generation, optimize prompts, prepare data, and evaluate performance with metrics aligned to business objectives such as productivity and user experience.

Evaluate cost, modality, latency, multilingual support, model size, complexity, customization, accountability, and bias or misinformation risks when selecting a pre-trained model; leverage retrieval augmented generation and fine-tuning.

Describe retrieval augmented generation (rag), a technique that blends information retrieval with language generation to deliver up-to-date, accurate answers from external knowledge bases.

Apply rag-powered retrieval augmented generation with AWS services to build a telecom chatbot that uses embeddings to access external knowledge, retrieve real-time data, and automate plan updates and orders.

Explore effective prompting strategies for generative AI, using instructions, context, input data, and output indicators to boost accuracy, reliability, and security while enabling domain specific knowledge and external tools.

Learn to modify prompts to unleash the full potential of generative AI, and tune inference parameters like temperature, top p, top k, maximum length, and stop sequences to guide creativity.

Apply best practices for prompt engineering to craft clear, concise prompts with context and directives. Start with a guiding question, specify the outcome, and break tasks into steps using templates.

Discover prompt engineering techniques such as zero-shot prompting, few-shot prompting, and chain-of-thought prompting, and how tuning and RLHF improve generative model performance.

Identify how foundational models respond to adversarial prompts and mitigate risks such as prompt poisoning, prompt hijacking, prompt injection, exposure, leakage, and jailbreaking.

Explore how data, algorithm, interaction, and amplification bias challenges responsible artificial intelligence, and learn to audit fairness, ensure diverse data, and address toxicity, hallucinations, intellectual property, and job disruption.

Explore the key dimensions of responsible AI: fairness, explainability, privacy and security, veracity and robustness, governance, transparency, safety, and controllability—to mitigate risks and build safer, more reliable AI systems.

Discover how responsible AI boosts trust and brand value, ensures regulatory compliance, mitigates risk, and drives competitive advantage, better decisions, and innovative, inclusive products.

Explore how AWS tools like SageMaker and Bedrock support responsible AI through model evaluation, bias detection, explainability, guardrails, monitoring, and governance.

Define the ai model use case to tailor performance and balance recall and precision for applications like missing persons search or virtual proctoring, while considering sustainability and ethics.

Balance data sets for responsible ai using SageMaker Clarify and SageMaker Data Wrangler to curate inclusive and diverse data aligned to the use case.

Learn how transparent and explainable models build trust and accountability, with tools like shap, lime, and counterfactual explanations, plus documentation, monitoring, and human oversight to ensure fair, auditable AI.

Discover AWS tools for transparency and explainability, including AI service cards, SageMaker model cards, and SageMaker clarify and autopilot, to support governance and explain model decisions.

Manage trade-offs between bias, variance, interpretability, safety, and controllability to build responsible, fair AI systems. Explore cross-validation, regularization, simpler models, and early stopping to reduce overfitting and improve generalization.

Explore how user centered design makes ai explanations clear and useful for diverse users. Apply amplified decision making, unbiased decisions, and human and ai learning to guide reinforcement learning with human feedback.

Explore AWS tools for AI governance, security, and compliance, identifying governance requirements and implementing controls to meet regulatory objectives and protect AI systems from risks.

Explore how AWS services enable governance and compliance in AI and generative AI workflows, including Config, Inspector, Audit Manager, Artifact, CloudTrail, and Trusted Advisor for secure, compliant operations.

Establish clear data governance strategies across the data life cycle to ensure data quality, privacy, and responsible AI. Define roles, monitor data, and enforce retention for compliant AI workflows.

Learn to implement governance strategies for generative AI, including policies, reviews, transparency, and training, and monitor performance, infrastructure, bias, and compliance with the five-scope security scoping matrix.

Discover security and privacy practices for ai systems, including threat detection, vulnerability management, infrastructure protection, and data encryption, plus the ten owasp vulnerabilities in large language models like prompt injection.

Explore how AWS security services like IAM, Macie, and WAF, within the shared responsibility model, protect AI systems from data breaches and adversarial attacks.

Explain data and model lineage to ensure traceability and reliability in AI, and document data provenance, licenses, and model cards for transparency.

Explore secure data engineering for generative AI, covering data quality metrics such as completeness, accuracy, and consistency, validations, privacy architectures including AWS privacy reference architecture, masking, encryption, and access controls.

Select the Northern Virginia region and create an S3 bucket named q business lab for demo files, then set up IAM Identity Center and a user with a password.

Create and deploy an Amazon Q business virtual assistant by configuring an app, selecting a native data retriever, and linking an S3 data source with run-on-demand synchronization.

Test a generative AI assistant in Amazon Q Business by logging in, querying revenue data with data source references from an S3 bucket, and uploading files.

Discover Amazon Rekognition, an AWS machine learning service that analyzes images and videos with AI. Learn label detection, facial analysis, text extraction, moderation, and real-time video analysis for media workflows.

Master Amazon SageMaker, a service that simplifies creating, training, and deploying ML models with SageMaker Studio, SageMaker Canvas, Data Wrangler, Feature Store, AutoML, and Jumpstart for real-time predictions.

Configure a human-in-the-loop workflow with Amazon Augmented AI in SageMaker, defining task types, confidence thresholds, selective keys, samples, templates, and worker options to validate AI results.

Learn how Amazon Comprehend uses machine learning and NLP to extract insights from text, including sentiment, key phrases, entities, PII, language, syntax, and topic modeling, from UTF-8 documents.

Launch Amazon Comprehend to analyze text with built-in analytics and custom models, extracting entities, key phrases, sentiment, and syntax, and perform topic modeling on S3 data.

Discover Amazon Kendra, a machine learning based intelligent search service that mimics a human expert to answer fact, descriptive, and keyword questions from indexed sources such as S3 and RDS.

Explore Amazon Lex, a machine learning service that builds conversational chat bots with automatic speech recognition and natural language understanding, exemplified by a hotel reservation bot.

Amazon Textract detects and analyzes text from documents, outputs extracted text and table structure, and enables analysis of receipts, invoices, and identity documents.

Explore Amazon Transcribe, an automatic speech recognition service that converts audio to text, with language customization, speaker identification, privacy filters, and custom vocabularies and language models.

Learn how Amazon Personalize builds ML personalization models from your data and trains e-commerce or video on demand recommenders with interactions data and promotional filters.

Learn about five AWS exam question types—multiple choice, multiple response, ordering, matching, and case study—and key strategies to identify keywords, eliminate distractors, and pace yourself for improved results.

Learn to schedule an AWS certification exam on the official AWS certification site, choosing in person or online, logging in, selecting date and time, and completing payment with required ID.