Course Code: bspaai
Duration: 35 hours
Prerequisites:

Experience with the Azure cloud platform

Course Outline:

Azure AI Training- Bespoke

Introduction to Machine Learning

  • What is Machine Learning (ML)
  • Some Classical ML Models
  • Evaluation of Models
  • Data Preprocessing
  • Neural Networks

Azure AI Fundamentals & Document Intelligence

  • Introduction to Azure AI services ecosystem
  • Azure AI services vs custom ML solutions
  • Hands-on: Creating Azure resources and configuring access
  • Document Intelligence (formerly Form Recognizer)
  • OCR and text extraction from supply chain documents
  • Custom document models for fashion industry documents
  • Hands-on: Building a document processing pipeline for garment specifications

Practice Project: Create an Azure Function that processes uploaded documents

(invoices, packing lists) and extracts structured data into a database

Computer Vision & Image Analysis

  • Azure Computer Vision capabilities
  • Image analysis and attribute detection
  • Custom Vision for fashion-specific attributes
  • Hands-on: Training a custom model to identify garment features
  • Image generation with DALL-E in Azure OpenAI
  • Prompt engineering for fashion-related images
  • Integrating generated images into product catalogs
  • Hands-on: Building an image generation API for product visualization

Practice Project: Develop a prototype that analyzes uploaded garment images

and automatically tags them with attributes (sleeve length, material type, number

of pockets)

Azure OpenAI & Language Models

  • Azure OpenAI Service overview
  • GPT models and capabilities
  • Prompt engineering best practices
  • Hands-on: Setting up Azure OpenAI and basic prompting
  • Fine-tuning language models for domain-specific tasks
  • Retrieval Augmented Generation (RAG) for supply chain knowledge
  • Hands-on: Building a RAG system with company documentation

Practice Project: Create a prototype that generates product descriptions and

specifications based on technical inputs

Building Custom Copilots

  • Copilot architecture and components
  • Orchestration patterns for AI services
  • Semantic Kernel framework
  • Hands-on: Building a basic Copilot framework
  • Integrating Copilot with existing VB.NET and Vue.js applications
  • Authentication and user context
  • Hands-on: Creating plugins for supply chain tasks

Practice Project: Develop a prototype Copilot that can automate common supply

chain tasks (e.g., generating a Bill of Materials, estimating production costs)

Azure Machine Learning & Production Deployment

  • Azure Machine Learning workspace
  • Training custom ML models for product costing
  • MLOps and deployment pipelines
  • Hands-on: Building and deploying a product cost estimation model
  • Integration strategies with existing .NET applications
  • Monitoring and observability
  • Security considerations for AI systems
  • Hands-on: End-to-end integration of AI services into the supply chain solution

Practice Project: Create a deployment plan for integrating the developed AI

capabilities into the production environment

Outlook - where to go from here