AWS Certified AI Practitioner AIF-C01 Updated 2026

Explore the fundamentals of AI and ML, including the application of foundation models and generative AI, and learn governance, security, and compliance for AI solutions in domain one.

Explore foundational AI and machine learning terms, including AI, machine learning, deep learning, and natural language processing, and see how generative AI and large language models generate text.

The instructor reassures you that you are not alone on your machine learning journey, invites questions in the Q&A, and praises your progress while encouraging module completion.

Define AI, ML, and deep learning, and explain large language models and NLP. Describe artificial neural networks trained on large data that learn patterns and generate new data.

Explore the core data types used in machine learning, including labeled and unlabeled data, tabular and time series data, and unstructured data such as text and images, with practical examples.

Explore the three types of machine learning: supervised, unsupervised, and reinforcement, distinguishing labeled versus unlabeled data, features, and targets, with practical examples.

Review the examination guide to map progress in domain one, identify key terminology, data types, and learning types, and note computer vision as learning advances.

Explore the fundamentals of classification, including binary, multi-class, and multi-label tasks, and review key algorithms—logistic regression, support vector machines, decision trees, random forests, and Naive Bayes.

Explore how to evaluate classification models by comparing predicted versus expected outputs using accuracy, precision, recall, f1 score, roc auc, and confusion matrix insights, guided by a decision tree example.

Explore how the confusion matrix reveals model performance beyond accuracy, detailing true positives, true negatives, false positives, and false negatives, and metrics like precision, recall, and F1.

Apply the confusion matrix to pick metrics for given scenarios. Use recall to minimize false negatives, precision to minimize false positives, and F1 score for balance.

Explore typical exam questions on machine learning types, evaluation metrics like accuracy, auc, roc, f1 score, and how to map scenarios to supervised, unsupervised, or reinforcement learning.

Discover how Amazon Transcribe, Amazon Translate, Amazon Comprehend, Amazon Lex, and Amazon Polly convert speech to text, translate languages, extract insights, and power conversational bots with text-to-speech.

Explore supervised regression methods, including linear, multiple and polynomial regression, and nonparametric approaches like support vector and tree-based models with kernel tricks and pruning to guard against overfitting.

Explore regression evaluation metrics, including mean squared error, rmse, mean absolute error, and r-squared, plus key topics like feature scaling, feature engineering, regularization, hyperparameter tuning, and cross validation.

Learn how data preprocessing improves model accuracy by addressing missing values, outliers, and feature scaling, then encode text with bag of words, tf-idf, and word embeddings.

Learn how cross-validation prevents overfitting and aids model generalization in machine learning by partitioning data into folds, training on k-1 folds, and averaging performance metrics.

Master hyperparameters and tuning methods—grid search, random search, Bayesian optimization, and genetic algorithms—to optimize performance and prevent overfitting in the machine learning workflow, including cross validation and evaluation metrics.

Explore unsupervised learning with unlabeled data to discover hidden patterns, reduce data dimensionality, and apply techniques like clustering, association rule mining, and anomaly detection for preprocessing and insight.

Explore fundamental machine learning concepts and unsupervised learning techniques, including clustering, dimensionality reduction, association rule mining, and anomaly detection, then preview deep learning architectures and their use cases.

Explore clustering techniques such as k-means, hierarchical, DBSCAN, and Gaussian mixture models. Cover dimensionality reduction with PCA, t-SNE, and autoencoders, time series analysis with trend, seasonality, cyclical, residual components.

Deep learning, a subset of machine learning, uses artificial neural networks to read data, extract features, and perform end-to-end learning for image recognition, natural language processing, and speech recognition.

Explore convolutional neural networks for computer vision, showing how convolution manipulates image pixels with a 2x2 colonel head moved across the image to enhance features like water bodies.

Explore inferencing types: batch, real-time, streaming, offline, and serverless—and their use cases. See how AWS SageMaker tools and ML ops enable production deployment, monitoring, and data prep.

Express gratitude for learners' dedication and commitment, invite feedback in the Q&A section, and encourage continuing momentum into domain two.

Discover how generative AI works and explore Amazon Bedrock in domain 2 of the AWS Certified AI Practitioner AIF-C01 Updated 2025 course.

Explore the fundamentals of generative AI, foundation models, and Amazon Bedrock to connect domain two and three for AWS Certified AI Practitioner exam, with insights on depth and real-time application.

Generative AI learns patterns from training data to generate novel text, artwork, music, and code, with AWS examples like ChatGPT text generation, Deepcomposer music, and Amazon Q code generation.

Explore foundation models in generative AI, trained on large diverse data sets to learn patterns. See how prompts and few-shot learning with transfer learning adapt models to tasks across modalities.

Explore large language models and foundation models, trained on vast data. Learn how they grasp language, predict next words, and generate coherent, human-like text across chats, translation, and storytelling.

Explore Amazon Bedrock, a fully managed, serverless AI service offering foundation models from Anthropic, Stability AI, and Amazon Titan FMS, with fine tuning and rag options.

Amazon Bedrock accelerates experimentation by enabling quick comparison of foundation models from multiple providers. Fine-tune with your data, build AI agents, and deploy scalable serverless applications on AWS infrastructure.

Explore Amazon Bedrock and the foundation-model concept, including base models, model inference, prompts, tokenization, and embeddings, all accessible through a single unified api.

Explore the Amazon Bedrock console, initialize the service, and learn how foundation models enable text generation, chat, summarization, image generation, and personalization with examples and code.

Explore how inference configuration in Amazon Bedrock controls output with temperature, top p, and response length, and practice hands-on in the Bedrock playground with Titan models.

Set up budgets in AWS billing and cost management to monitor spend, receive email notifications, and enforce zero spend or monthly budget thresholds.

Explore Amazon bedrock providers and models, using the built-in playground to try them, and request model access for providers like eye to eye labs, Amazon Anthrop ic, Meta, and Stability AI.

Explore how Bedrock enables fine tuning and continued pre-training on foundation models to create custom models; prepare prompt-completion data, set hyperparameters, and import models or use the playground.

Review the cost factors of generative AI on Amazon Bedrock, focusing on input and output tokens, and prepare for the AWS AI Practitioner exam with docs and optional practice tests.

Learn how to create, test, and publish reusable prompts in the Amazon bedrock prompt management tool, configure inference parameters like top-p and temperature, and deploy prompts in production.

Explore how knowledge bases enable retrieval augmented generation by connecting external data to foundation models, using embeddings and vector stores to answer from S3, SharePoint, or Confluence.

Learn how Titan embed text v1 converts text to 1536-dimension vectors and stores them in a knowledge store with Bedrock, OpenSearch, or MongoDB for enterprise retrieval.

Explore RAG vector databases with OpenSearch, scalable index management, and k-nn search; compare DocumentDB, Aurora, RDS PostgreSQL, and Neptune for vector embeddings and data sources like S3.

Explore Amazon Bedrock agents and prompt flows that use foundation models to orchestrate API calls, data sources, and embedding-based knowledge bases for automated flight ticket booking.

Amazon Bedrock guardrails enforce safeguards to protect personal data, block sensitive prompts and prompt attacks, and filter content, while testing with models and enabling watermark detection for Titan image generation.

Explore how Amazon Bedrock manages inference with provisioned throughput for dedicated capacity, batch and real-time inference, and cross-region inference to ensure scalable, resilient AI deployments.

Explore Amazon Bedrock configurations, including Bedrock Studio and workspace access via IAM, pricing for input/output tokens, and SDK or API access, plus security with encryption, KMS, VPC, and embeddings.

Explore Amazon Bedrock security with IAM and identity-based policies, monitor model invocations using CloudWatch, CloudTrail, and EventBridge, and review the syllabus for generative AI concepts.

Learn the architecture of a prompt, including the three pillars—system message, context, and user input—and how prompt engineering shapes effective interactions with generative AI.

Explore prompt engineering techniques, including zero shot prompting, few short prompting, chain of thought prompting, role prompting, and instruction based prompting, plus prompt response history and context.

Explore data augmentation for language models by expanding training data through paraphrasing, back translation, and synonym replacement. Recognize why data preparation drives LM training.

Discover prompt engineering, reinforcement learning with human feedback (rlhf), and low rank adaptation (LoRA) for efficient, human-aligned fine tuning of large language models across pre-training and fine-tuning.

Apply AB testing to LLMs by comparing version A and version B, collecting metrics, analyzing data, and selecting the best model while addressing bias and using human evaluation.

Evaluate large language models and RAG applications using task-specific metrics such as accuracy, perplexity, blue/rouge, and retrieval, augmentation, and generation quality measures.

Explore techniques to optimize output generation from large language models, addressing latency and resource constraints through pruning, quantization, knowledge distillation, batching, caching, and hardware acceleration.

Master zero shot testing and benchmarking for large language models, evaluating tasks without prior training, using datasets like glue, superglue, squad, and wmt to measure progress.

Explore hallucination in large language models, where outputs seem plausible yet are factually incorrect. Learn causes like data bias and grounding gaps, and apply guardrails and verification to boost trust.

Master domain 1–3 fundamentals of AI, ML, and DL, including computer vision, natural language processing, bias, fairness, and the ML lifecycle with SageMaker and Bedrock concepts.

Explore Amazon Q, a Gen AI powered assistant that accelerates software development by leveraging internal data and integrates with QuickSight and Connect to generate code, tests, and multi-step workflows.

Explore Amazon Bedrock's Party Rock playground for gen AI apps using SageMaker Jumpstart's NLP, foundation, and vision models, and review Amazon Q pricing options.

Explore guidelines for responsible AI and generative AI, implement guardrails to filter input and outputs, and use SageMaker clarify and model monitor to detect bias and ensure safety.

Secure ai systems with IAM policies, encryption, and Amazon Macie to identify PII in S3, while managing data lineage, cataloging, and governance through SageMaker model cards and CloudTrail.