Posts and Ideas about AI, Machine Learning, and Cloud Computing
1 videos
Grouping together videos that build on the NFLVerse data set
5 videos
These videos cover configuration, use and thoughts around the NVIDIA DGX Spark platform
4 videos
A list of videos covering various Azure configuration and administration topics
11 videos
23 videos
6:26
This video walks through the process of setting up Microsoft Fabric database mirroring from Google BigQuery, including an automated script for configuring the GCP side. You'll learn how to create a mirrored database in Fabric, set up near real-time replication, and validate the data synchronization between BigQuery and Fabric.
5:20
This video walks through how to use Microsoft Fabric shortcuts to access Google BigQuery Iceberg tables, automating the setup process on the Google Cloud side and configuring direct SQL queries within Fabric without copying or moving data. You'll learn about the automation script that handles service account creation, IAM permissions, and required BigQuery table access, followed by setting up connections and creating shortcuts in Fabric for seamless data querying.
33:16
This video walks through the process of building an end-to-end analytics reference architecture for the National Football League (NFL) using Microsoft Fabric. Viewers will learn how to download and process NFL data, create a Fabric workspace and Lakehouse, deploy Power BI semantic models, and develop AI-powered analytics agents, all while exploring key concepts like data modeling and agent evaluation workflows.
7:29
This video walks through the process of exposing on-premises SQL Server data for natural language querying using Microsoft Fabric Mirroring and OneLake. You'll see how to set up Change Data Capture (CDC) and the On Premises Data Gateway, mirror specific tables into OneLake, and use a Fabric Data Agent to enable users to ask plain English questions that are then processed by Fabric without manual SQL intervention.
17:22
This video walks through configuring a Microsoft Fabric workspace with a Workspace-scoped Private Link, demonstrating how to restrict access to an Azure Virtual Network (VNet) and build a secure analytics architecture that combines strong network isolation with flexible reporting access. Viewers will learn about the setup process, including creating virtual networks, subnets, private endpoints, and deploying a jump box VM for testing private connectivity.
4:31
This video walks through the process of building a Microsoft Fabric ontology using Python code within a Fabric notebook. You'll learn how to define entities, relationships, and properties using Python, as well as automate the deployment process via the Fabric REST API.
3:09
This video walks through the setup of a real-time streaming pipeline in Microsoft Fabric, demonstrating how telemetry data from trucking equipment is processed to automatically capture critical operational events. Viewers will see how telematics events are streamed into Fabric, filtered for mechanical faults, stored in a Lakehouse table connected to an ontology model, and used to trigger automated workflows, all while visualizing live event movement through the system.
1:55
This video walks through the integration of a Microsoft Fabric Data Agent with a Microsoft Foundry Agent to create a versatile multi-source AI assistant. Viewers will learn how to combine structured enterprise data from a Fabric Ontology with unstructured documents using Azure AI Search, enabling the agent to answer complex questions and support natural language queries across various knowledge sources.
4:23
This video walks through building a secure, production-ready web application using Microsoft Fabric, Fabric Ontology, and Microsoft Foundry Agents to integrate map-based visualizations and generative AI experiences within a custom React app. The walkthrough covers the architecture of the solution, including the React frontend with Next.js, FastAPI Python backend, and various Azure services like AI Search and Maps.
41:08
This video walks through the process of designing and building an ontology in Microsoft Fabric, covering key aspects such as defining entity types and relationships, applying best practices for relationship naming, and implementing these concepts step-by-step within the platform. Viewers will gain insights into what an ontology is and how it can be used for data analysis in a business context.
9:04
This video walks through the basics of graph data analysis and demonstrates how Microsoft Fabric's new graph capabilities can be used to build and query graphs, focusing on practical examples using customer and order data. Viewers will learn about graph models, nodes, edges, and relationships, as well as how to use GQL for querying graphs within a Fabric Notebook.
10:58
This video walks through the introduction and initial setup of Microsoft Foundry Local, a new local inference engine designed for running AI models on Windows, macOS, or Windows Server 2025. It covers installation steps, testing local inference performance, and building applications using the OpenAI-compatible SDK, providing a comprehensive guide for developers looking to integrate secure and high-performance AI into their workflows.
20:04
This video walks through the process of building a custom MCP (Model Component Provider) tool for use with Azure AI Foundry Agents, specifically focusing on integrating a Python-based server that exposes GraphRAG tools via HTTP. You'll learn how to design and implement this server as an Azure Function, deploy it, and connect it to an Azure AI Foundry Agent to enable querying of a Cosmos DB Gremlin graph without requiring knowledge of the underlying data source.
18:10
This video walks through the concepts and implementation of a GraphRAG application, explaining what graph databases are and how they enhance language model retrieval. It also contrasts GraphRAG with vector RAG to highlight their respective strengths and applications. The walkthrough includes a detailed code review of a full-stack GraphRAG app, providing insight into its backend, graph database integration, LLM usage, and user interface.
2:51
This video walks through a quick demo where LinkedIn profile data is fed into Gemini Pro via the new Google Nano Banana Pro to generate an infographic. The viewer will see how well the model handles complex typography, consistent fonts, and structured layouts, ultimately assessing its accuracy and readability in presenting professional information.
![Vectorizing Images with LLMs [Full Stack Deep Dive]](https://i.ytimg.com/vi/BQqmkyBAU_I/maxresdefault.jpg)
13:04
This video walks through building a complete end-to-end image vectorization and similarity search system using modern AI tools, including multimodal LLMs and GPU-powered infrastructure. You'll learn how to generate image embeddings with an OpenAI model, store vectors efficiently with ChromaDB, and integrate LLM microservices for semantic search, all while exploring the architecture and full demo of the solution.
37:19
This video walks through the full workflow of fine-tuning the Phi-4-Mini-Instruct model using LoRA and PyTorch, covering how to format custom training data, code a fine-tuning script, run jobs in a Docker container, merge fine-tuned adapters, and quantize the model for efficient local or edge device operation. Viewers will gain hands-on experience with these techniques applicable to any CUDA-based environment.
9:39
This video walks through the three key patterns of Agentic AI—Chat Agents, Workflow Agents, and Autonomous Agents—and provides practical code examples for each. You'll learn how reasoning models and tools work together in agentic systems and when to apply each pattern to solve real-world problems.
5:10
This video walks through the process of setting up OpenAI’s GPT-OSS-20B locally using Docker and Open WebUI, allowing viewers to experiment with a large language model on their own hardware. You'll learn how to create a Docker container, set up NVIDIA Sync for macOS, and download and run the model on various types of machines, including desktop-class systems like the NVIDIA DGX Spark.
5:47
This video walks through the NVIDIA DGX Spark, detailing its hardware specifications and value proposition for AI system development. It explores how this data-center-grade AI lab fits on a desk but is tailored to specific users, providing insights into whether it meets your needs for building, testing, and deploying AI systems from prototype to production.
9:37
This video walks through the features of OpenAI's newly launched Atlas browser, exploring its capabilities beyond traditional web browsing. The presenter tests various functions, including page summarization and trip planning, to determine if Atlas lives up to its promise as a disruptive technology in the digital landscape.
18:06
This video provides a step-by-step walk-through of configuring an Azure Entra ID application registration to enable single sign-on (SSO) for accessing a Snowflake database from a React web application. Viewers will learn how to set up the necessary connections and integrations, with techniques applicable to other applications that can leverage Entra via MSAL or REST connections.
6:23
This video walks through the process of integrating Entra OAuth authentication into a Streamlit web application, covering steps from creating an Entra app registration to running the application and handling user consent. Viewers will see how to configure MSAL and explore the code behind the implementation.
7:27
This video walks through the process of integrating a Microsoft Fabric Semantic Model query into a custom web application built with React and TypeScript, covering authorization, token fetching, and querying semantic models using the Power BI REST API. Viewers will learn how to add necessary permissions, fetch access tokens, and implement code within their React app to submit queries and display results.
31:03
This video walks through the process of setting up an App Registration in Microsoft Entra for a React application and integrating it with MSAL (Microsoft Authentication Library) to handle authentication flows, including login, logout, token refresh, and user assignment. Viewers will learn how to configure the app registration, set authorized redirect URLs, restrict access to specific users or groups, and implement essential login functionalities from scratch in a new React project.
8:38
This video walks through the process of deploying an Azure Web App using a DevOps pipeline, including setting up a custom app registration and service profile secret. Viewers will learn to configure steps such as creating an app registration in Entra, generating a client secret, authorizing a service principal, establishing a DevOps Service Connection, and creating and running the deployment pipeline.
18:16
This video walks through the process of setting up a local Windows machine with an Nvidia GPU to fine-tune large language models (LLMs) for specific scenarios. Viewers will learn how to install necessary tools and components, create a Python environment, and fine-tune an LLM using a training notebook.
7:03
This video walks through the process of using Fabric mirroring to replicate data from a Snowflake Data Warehouse into a Microsoft Fabric Data Warehouse, focusing on enabling mirroring with Change Data Capture for near real-time data replication. Viewers will see how to create and manage mirrors, select tables, monitor mirror status, and review and add new data.
5:27
This video walks through the process of using a Fabric pipeline to import data from a Snowflake Data Warehouse into a Lakehouse Delta Table, demonstrating how to create a pipeline, establish a connection with Snowflake, select the appropriate data, and run the pipeline to ensure successful integration.
6:23
This video walks through the process of exporting a machine learning model developed in Microsoft Fabric, importing it into an Azure Machine Learning workspace, and deploying it to a real-time inference endpoint. Viewers will learn how to set up and test this deployment, ensuring their models can provide real-time predictions efficiently.
8:16
This video walks through the process of selecting pre-trained models from Azure ML's Model Catalog and deploying them on Azure's computing services, culminating in accessing these models via a REST API. Viewers will learn how to create a resource group, set up a workspace, and test an endpoint.
13:30
This video walks through the process of creating and installing custom Python libraries in a Microsoft Fabric Lakehouse environment, demonstrating how to package code for reuse across notebooks. Viewers will learn how to create a Python package that fetches secrets from Azure Key Vault and install it within their Fabric workspace.
7:09
This video walks through integrating Azure AI sentiment analysis models into a Fabric Jupyter notebook environment, demonstrating how to enhance Data Lake tables with user sentiment data. Viewers will learn how to leverage these models for uncovering deeper insights in unstructured data sets.
8:43
This video walks through using Azure OpenAI Services within a Fabric Jupyter notebook to automatically summarize large sets of semi-structured user reviews. It also covers prompt engineering, secure secret storage in Azure Key Vault, and data handling with Delta tables in a Fabric Lakehouse environment.
15:31
This video walks through the process of building a Document Intelligence solution using Azure AI within Microsoft Fabric. It demonstrates how to ingest scanned document data into a Data Lake solution by leveraging Spark Notebook and Synapse ML for text extraction.
5:42
This video walks through the process of integrating Azure AI Translation into a Fabric Data Lake solution using Synapse ML features. Viewers will learn how to apply Language AI Services within Python Jupyter notebooks stored in a Fabric workspace for data enrichment and transformation.
9:35
This video walks through the process of calling Azure AI Document Intelligence services using the REST API, enabling viewers to extract data from images and PDF scans of forms directly via API calls.
9:01
This video walks through setting up and using Git repositories with Microsoft Fabric Workspaces, covering configuration, syncing notebooks and other content, and how to recover unintended changes. You'll learn to connect a Git repository to your workspace, commit changes, and revert modifications as needed.
9:55
This video walks through the process of building a fully functional Gemini Chatbot using Streamlit, starting with setting up Anaconda and installing dependencies before initializing the Gemini SDK and creating a user prompt interface. The viewer will see how to run the bot and observe its interactions in real-time.
18:58
This video walks through creating a custom Generative AI Microsoft Teams copilot using Copilot Studio, covering the process from scratch to deployment. Viewers will learn how to integrate web site and PDF documentation as grounding knowledge, add custom topics with branching logic, and publish the copilot to a Teams channel.
4:26
This video walks through the process of displaying a Power BI report directly within a Jupyter notebook hosted on Microsoft Fabric using the Power BI Python client. Viewers will learn how to authenticate, import necessary dependencies, and create a report instance in just a few steps.
10:58
This video walks through the process of setting up an Azure AI Search service with integrated vector embeddings to index documents for use in Retrieval Augmented Generation (RAG) within generative AI Q&A solutions. It then demonstrates creating a Streamlit Python application that interacts with an OpenAI large language model, utilizing the indexed data from Azure AI Search as the RAG data source.
2:35
This video walks through the process of creating images with Bing Image Creator, which leverages DALL-E technology, and then enhancing those images in Photoshop using the Generative Fill feature.
9:54
This video walks through how to use Fabric Semantic Link to integrate Power BI as a data source for developing machine learning models. Viewers will see a step-by-step guide on connecting to Power BI measures, querying the data, and training an ML model within Microsoft Fabric.
4:37
This video walks through the process of creating a machine learning model using IBM Watson Studio and Jupyter notebooks, covering essential steps such as logging into IBM Cloud, setting up a project, creating a Jupyter notebook, and training a model.
14:39
This video walks through the process of building a generative AI chatbot that leverages Retrieval Augmented Generation (RAG) to query OpenAI's Large Language Model using a vector database. Viewers will learn how to set up this system, encode custom data, and test both the vector database and the resulting chatbot functionality.
6:26
This video walks through the process of moving a local Jupyter notebook from Anaconda to Microsoft Fabric Data Science, using a dataset on SpaceX launches to predict successful missions. Viewers will learn how to leverage Fabric Data Science without needing prior knowledge of developing machine learning models.
6:52
This video walks through the process of integrating private data into Microsoft Azure OpenAI Service using Retrieval Augmented Generation (RAG) techniques, including uploading PDF files and creating an OpenAI deployment. Viewers will learn how to set up a RAG index and test both sandbox and Python environments for generating AI queries that incorporate their own data.
19:11
This video walks through the process of deploying a machine learning model to a REST endpoint using Microsoft Azure Machine Learning, covering all necessary steps from creating resources and launching ML Studio to training the model and registering it before finally deploying the inference endpoint. Viewers will learn how to integrate their trained model into production applications by making predictions available via HTTP requests.
13:48
This video walks through the process of exporting AI Vision models from Azure cloud services and integrating them into iOS applications using CoreML. Viewers will learn how to test, export, and run these models locally on iOS devices, enhancing their understanding of edge computing in AI applications.
10:50
This video walks through the process of creating an Azure AI Custom Vision object detection model to identify packaging defects in products. Viewers will see how to set up and train a custom vision project, from creating necessary Azure resources to importing images, tagging them, and testing the trained model for accuracy.
11:34
This video walks through the process of using Azure AI Vision to generate descriptive metadata for images, and then integrating this data into a Power BI analysis solution. Viewers will see a live demonstration of creating an enriched image dataset and how it can be visualized in Power BI.