Azure AI services complete guide - Covers AI-102 Cert

In this lesson, we will see Azure Content Moderator, a tool designed to help keep your applications safe and appropriate for users. We'll discuss how it can automatically detect and filter offensive language, harmful imagery, and unwanted content.

In this lesson, we'll learn how to effectively moderate text content using Azure Content Moderator, ensuring that all user-generated text remains safe, respectful, and appropriate.

In this lesson, we’ll explore how to moderate image content using Azure Content Moderator and filter out any inappropriate or unwanted content.

In this lesson we will discuss about Azure AI Content Safety which is the new name for Azure Content Moderator. We'll also look at how to call the Content Moderator API using Python code and get results for text and image moderation.

In this lesson, we’ll discuss the Content Safety Blocklist feature in Azure Content Moderator, which helps filter out unwanted or offensive text based on the list of custom terms we provide as a list.

In this lesson, we will discuss how to analyze images using Azure AI Vision to extract insights and detect important features. Also to generate tag for images, generate caption, dense captions etc.

In this lesson, we will discuss Landmark Detection with Azure AI Vision, which enables the recognition of famous landmarks and locations within images.

In this lesson, we will explore how to use Azure AI Vision for detecting brands and logos in images, extracts text from images using OCR (Optical Character Recognition) and automatically generate tags for images.

In this lesson, we will cover image classification using Azure's Custom Vision service. This feature allows you to train models to categorize images into specific classes, helping applications automatically identify and organize visual content based on predefined labels.

In this lesson, we will cover more details on image classification using Azure's Custom Vision service. Wee will also have a look at the custom vision studio.

In this lesson, we will explore evaluation metrics and object detection within Azure's Custom Vision service. We'll discuss how to measure model accuracy with metrics like precision, recall, and mean average precision. Additionally, we'll learn about object detection capabilities, which enable models to detect and locate multiple objects within an image, providing bounding boxes.

In this lesson, we will learn about the Conversational Language Understanding model provided by Azure AI Language. We'll discuss how this model can be utilized to build natural and effective conversational agents. We will also learn about entity, intent and utterances.

In this lesson, we will demonstrate the capabilities of the Conversational Language Understanding model within Azure AI Language.

In this lesson, we will explore Azure AI Language to analyze text. We'll cover extracting summaries, detecting languages, linking entities, extracting key phrases, and identifying named entities.

In this lesson, we will explore how to convert text into speech using the Azure AI Speech Service.

In this lesson, we will focus on converting spoken language into text using the Azure AI Speech Service. Additionally, we will explore the translation capabilities of the service, allowing you to convert spoken language from one language to another in real-time.

In this lesson, we will discuss SSML (Speech Synthesis Markup Language) and its role in enhancing speech generation with the Azure AI Speech Service.

In this lesson, we will explore Custom Speech using the Azure AI Speech Service. We will see how to train a baseline model with our own data and improve the model. We will also have a look at the speech studio.

In this lesson, we will explore CI/CD concepts in the context of MLOps (Machine Learning Operations). We will discuss how continuous integration and continuous deployment practices can be applied to machine learning projects, focusing on automating the deployment of models.

In this lesson, we will focus on deploying a Custom Speech model using CI/CD (Continuous Integration and Continuous Deployment) practices. We'll discuss the steps involved in automating the deployment process, including version control for your models, integration with CI/CD pipelines, and best practices for testing and validation.

In this lesson, we will deal with the Azure AI Translator service, a powerful tool for real-time language translation. Azure AI Translator enables you to translate text and documents across multiple languages effortlessly.

In this lesson, we will explore Azure AI Document Intelligence, focusing on its prebuilt models designed to automate the extraction of information from various document types. We will use invoice, identity document, layout model in this demo.

In this lesson, we will be creating a Custom Document Intelligence model using Azure AI. We will cover how to train and fine-tune models to extract specific information tailored to your business needs, enabling you to process unique document types effectively.

In this lesson, we will explore the key concepts of Azure AI Search, Indexing and concepts like filterable, searchable, facetable, sortable etc.

We will be seeing a demo on Azure AI Search in this lesson which includes creating a project, uploading and indexing the document and querying the data.

In this lesson, we will discuss how to deploy OpenAI's GPT model and effectively utilize prompts and parameters to customize its responses. Additionally, we will cover the implementation of content filters to ensure safe and appropriate outputs in various applications.

In this lesson, we will explore how to fine-tune OpenAI models to enhance performance using private data. We will also discuss the capabilities of DALL-E in generating images from textual descriptions.

In this lesson, we will discuss the Azure AI Video Analyzer, focusing on its capabilities to extract insights from video content through advanced analysis techniques.

In this lesson, we will explore Azure AI Vision's Spatial Analysis capabilities, which enable the understanding of physical space through video data. We will discuss how it can track objects, analyze movements, and provide insights into interactions within an environment, enhancing applications such as smart surveillance and retail analytics.

In this lesson, we will discuss how to effectively manage, monitor, and secure Azure AI services by utilizing API keys, assigning roles, and analyzing logs for performance insights. Additionally, we will explore best practices to control access to AI resources by using virtual networks to enhance security.

In this lesson, we will discuss how to plan and implement a container deployment using Docker for Azure's Text-to-Speech service.

In this lesson, we will explore how to build a Custom Question Answering solution using Azure AI Language. We will upload document, add alternation QnA pair, implement multi-turn conversation and deploy the chatbot.

In this lecture, we will explore Azure AI Foundry, Microsoft's end-to-end platform for building, fine-tuning, and deploying AI applications at scale. Here’s what we’ll cover:

Creating a Hub & Project in AI Foundry – Learn how to set up an AI hub and initialize projects to organize your AI workflows efficiently.

AI Models in Foundry – Discover the different pre-built and customizable AI models available, including Azure OpenAI, open-source, and Microsoft-developed models.

Fine-Tuning Models in the Playground – Understand how to customize AI models for specific use cases by adjusting parameters.

In this lecture we will be discussing about agents, what are the use cases and how to deploy agents in Azure AI Foundry.

In this lecture, we will see a demo on agents which uses an action tool, code interpreter and getting things done. We will be creating a graph diagram from a .csv file using agents.

In this lecture we will see agents executing local functions.

Prompt flow is a development tool designed to streamline the entire development cycle of AI applications powered by Large Language Models (LLMs). In this lecture we will discuss about different types of prompt flow, see prompt flow in action along with prompt templates.

In this lesson we will be discussing about tracing in Azure AI Foundry, which is used to monitor and visualize the execution flow of AI applications.

"RAG" stands for Retrieval-Augmented Generation. It's a technique that improves the output of large language models (LLMs) by incorporating information retrieval before generating a response. Essentially, RAG allows LLMs to access and reference external knowledge sources (like databases or documents) to supplement their own training data, leading to more accurate, relevant, and up-to-date answers. In this lesson we will see in detail the concepts of RAG and it's demo.

Link to code for RAG:

https://github.com/Azure-Samples/azure-search-openai-demo

Semantic Kernel (SK) is an open-source software development kit (SDK) that simplifies the integration of large language models (LLMs) and other AI services into applications. In this lesson we will be discussing about langchain and semantic kernel and will see a chatbot built using semantic kernel.

Link to the code used in the demo:

https://github.com/microsoft/semantic-kernel/blob/main/python/samples/getting_started/00-getting-started.ipynb

Link to the documentation mentioned in the video:

https://techcommunity.microsoft.com/blog/educatordeveloperblog/llm-based-development-tools-promptflow-vs-langchain-vs-semantic-kernel/4149252

In this lesson we will be discussing about AutoGen a framework for building AI agents and applications. Also, we will see the demo of a simple chess game that you can play with an AI agent.

Link to Code:

https://github.com/microsoft/autogen/tree/main/python/samples/agentchat_chess_game

MCP (Model Context Protocol) is an open-source protocol that enables AI agents to access external data and tools in a consistent and standardized way. It allows agents to access information and perform actions by interacting with MCP-compatible servers, essentially creating a standardized way for agents to "talk" to external resources. In this lesson we will be seeing a demo on MCP.