AI-900: Microsoft Certified: Azure AI Fundamentals

Explore artificial intelligence fundamentals: perception, reasoning, learning, and natural interactions, and see AI workloads like machine learning, computer vision, natural language processing, and generative AI.

Explore Microsoft’s six responsible AI principles: fairness, reliability and safety, privacy and security, inclusiveness, transparency, and accountability—and see how Azure Machine Learning helps detect biases and ensure safe, trustworthy AI.

Discover how machine learning uses training data, features, labels, and algorithms to build models that identify patterns and make predictions and decisions.

Explore supervised and unsupervised machine learning, including regression and classification (binary, multi-class, multi-label), and clustering, with real-world examples like sales, property prices, and customer segmentation.

Split data into training and validation sets, train the model to learn features and labels, and evaluate performance with metrics like accuracy or mean squared error, iterating to improve.

Explore deep learning, a powerful subset of machine learning, using deep neural networks with multiple layers, activation functions, softmax, and backpropagation to tackle classification, regression, NLP, and computer vision.

Explore automated machine learning in Azure Machine Learning Studio to train, deploy, and test a regression model on bike rental data, with a real-time endpoint for predictions.

Discover Azure AI vision capabilities to analyze and interpret images and videos, empower businesses with meaningful insights, explore its services, and learn the face detection feature for real-world applications.

Explore Azure AI Vision Services 4.0, generally available, for image analysis with OCR, people detection, and brand recognition, and train robust models using diverse images stored in Azure storage.

Discover how computers interpret images as numerical pixel data, using grayscale and RGB channels, and how convolutional filtering with kernels reveals edges for object detection and facial recognition.

Explore how grayscale and color images are processed using kernels and convolution to apply filters, revealing edges with Laplace filters and enhancing image contours.

Explore how convolutional neural networks, a deep learning model for computer vision, extract features with filters to classify images and detect and segment objects.

Explore Azure AI Vision multimodal models, contrast keyword search with embedding and vector search, and learn how Florence enables image analysis tasks like classification, object detection, and captioning.

Learn how Azure AI Vision image analysis 4.0 extracts meaningful insights from images with OCR, object detection, dense captioning, and people detection, compared to 3.2.

Explore how Azure AI vision analyzes store images using Vision Studio to generate captions, tags, and object and face detection, enabling smarter customer service.

Explore optical character recognition (OCR) with Azure AI Vision to extract printed and handwritten text from images and documents, using cloud OCR for images and Document Intelligence for text-heavy documents.

Explore optical character recognition with Azure AI vision using Vision Studio to extract text from images without writing code, converting it into searchable, editable data for automated processing.

Explore natural language processing (NLP) and its use in Azure AI services, recognize when NLP and conversational AI apply, identify services that enable text analysis to extract insights from data.

Natural language processing enables machines to understand, process, and generate language, using preprocessing, tokenization, stemming, lemmatization, and stop-word removal to train models for sentiment analysis, translation, and summarization.

Explore how Azure enables natural language processing to extract sentiment, topics, language, key phrases, and document classification, then apply semantic search and knowledge graphs for real-world automation.

Azure language, speech, and translator services analyze text with entity recognition, sentiment, key phrase extraction, and classifications, enable speech to text, text to speech, and real-time translations.

Explore Azure AI language services with Language Studio to build and train a knowledge base for question answering, deploy it, and enhance customer interactions and automations.

Learn how CLU, Azure AI language's conversational language understanding, interprets natural language input to predict user intent and extract entities, then train, deploy, and optimize models for real-world apps.

Explore how Azure AI Language and Language Studio power a home automation model that interprets spoken or text commands, training intents, entities, and deploying a live demo.

Explore how Azure AI Speech Service converts speech to text and text to speech in real time within Speech Studio, enabling automated transcriptions and lifelike voice synthesis for AI-driven experiences.

Explore how the Azure AI translator translates text and speech across languages in real time via the Azure portal, with language detection and access to keys, endpoints, and web APIs.

Analyze customer feedback with sentiment analysis and key phrase extraction using Azure AI Language Studio to classify text and extract insights from hotel reviews.

Explore Azure AI Document Intelligence and Azure AI Search to unlock data insights, streamline document processes, and improve decision making through knowledge mining of unstructured data.

Explore how Azure AI Document Intelligence automates data extraction from business documents, using document analysis, prebuilt and custom models, across cloud, on-premises, or edge deployments to streamline workflows.

See how Azure AI Document Intelligence goes beyond OCR to extract structured data from forms and receipts with prebuilt models, producing key values, tables, and JSON-ready output.

Explore how Azure AI Knowledge Mining uses AI-driven processing to ingest, enrich, and index unstructured data from documents and notes, enabling fast, AI-powered enterprise search for actionable insights.

Explore how Azure AI search ingests, indexes, and queries structured and unstructured content with full text and vector search, enabling AI powered enrichment and semantic ranking for retrieval augmented generation.

Explore how Azure AI search transforms raw data into an optimized index via a five-stage pipeline: data source, indexer, document cracking, AI enrichment, and indexing.

Build a knowledge mining solution with Azure AI search to extract, enrich, and organize customer reviews, enabling fast search and actionable insights from the data.

Explore the fundamentals of generative AI, how language models generate text and images, and apply these concepts in Microsoft Copilot and Azure AI Foundry, while embracing responsible AI practices.

Generative AI is a specialized branch of artificial intelligence that creates original content from a natural language prompt, across text, images, audio, video, and code.

Explore the transformer architecture behind modern language models, including bert and gpt, and how tokenization, embeddings, and self-attention enable coherent, context-aware text generation.

Learn how tokenization breaks text into tokens for transformer models, including words, word parts, and punctuation, assigns ids to standardize inputs, and enable efficient learning and generalization in language models.

Explore how embeddings encode semantic meaning as multi-dimensional vectors, compare semantic similarity using cosine similarity, and respect 8192 token input limits, 2048 inputs per request, and 350k tpm deployment limits.

Explore how transformer attention builds context for language models by using self-attention in encoder and decoder blocks, with multi-head attention to predict the next token from token embeddings.

Leverage foundation and pre-trained language models on Microsoft Azure with the Azure OpenAI service, using a centralized model catalog and fine-tuning to tailor AI for industry-specific tasks.

Explore how generative AI assistants like Copilot boost productivity with AI powered suggestions, content creation, and automation across M365 and Dynamics 365, including pre built, extended, and custom AI agents.

Craft clear, goal-driven prompts and apply prompt engineering to generative AI, grounding responses with real data, defining formats, and iterating based on past results.

Explore Copilot studio for low-code creation of conversational AI inside Microsoft 365, and Azure AI Foundry for full control, fine-tuning, and prompt flow to build tailored agents.

Explore, build, test, and deploy scalable generative AI applications on an enterprise-grade platform with Azure AI foundry, guided by responsible AI practices.

Azure AI Foundry unifies models, catalogs, and prompt flow with OpenAI service and AI services in a single workspace, enabling developers to build, deploy, and scale AI solutions.

Explore azure ai foundry, hubs and projects to organize data, artifacts, and compute resources, while using prompt flows, foundation models, and governance with RBAC.

Explore a four-stage framework for responsible generative AI: identify harms, measure harm presence, mitigate across model, safety, grounding, and experience layers, and operate responsibly after deployment.