AI-102 Exam Preparation (Part 1): Prerequisities

This a breakdown of concepts to know for each of the prerequisites for AI-102, so you can quickly assess and fill any gaps before diving deep into the exam content.

---

✅ 1. AZ-900-Level Understanding of Azure (Cloud Fundamentals)

🌐 General Cloud Concepts

• What is cloud computing? (IaaS, PaaS, SaaS)
• Public, private, and hybrid clouds
• Benefits: scalability, elasticity, fault tolerance, high availability

☁️ Core Azure Services

• Compute: VMs, App Services, Containers
• Networking: VNets, Load Balancer, VPN Gateway
• Storage: Blob, Table, Queue, File storage
• Databases: Cosmos DB, Azure SQL Database

🧩 Azure Architecture & Management Tools

• Resource Groups
• Azure Resource Manager (ARM)
• Azure Portal, Azure CLI, PowerShell
• Regions, availability zones, and paired regions

🔐 Security, Compliance, Identity

• Azure Active Directory (Azure AD)
• RBAC (Role-Based Access Control)
• Network Security Groups (NSGs), Firewalls
• Compliance certifications

💰 Pricing and SLA

• Azure pricing calculator
• Total Cost of Ownership (TCO) and cost management
• SLA (Service Level Agreements) basics

---

✅ 2. AI-900-Level Understanding of Azure AI

🧠 Fundamentals of AI Workloads

• What is AI vs ML vs DL?
• Supervised vs unsupervised learning
• Use cases: vision, language, prediction, conversation

🤖 Azure AI Services Overview

• Cognitive Services
  • Vision: Computer Vision, Face, Custom Vision
  • Language: Text Analytics, Translator, Language Understanding (LUIS/CLU)
  • Speech: Speech-to-text, Text-to-speech
  • Decision: Personalizer, Content Moderator
• Azure Machine Learning
  • Designer, AutoML, notebooks
• Conversational AI
  • Azure Bot Service
  • Bot Framework and Bot Composer

🔒 Responsible AI Principles

• Fairness, Reliability, Privacy, Inclusiveness, Transparency, Accountability

⚙️ Deployment and Integration Basics

• How Cognitive Services are deployed (containers vs cloud)
• Using REST APIs and SDKs to access services
• Common real-world AI application scenarios

---

✅ 3. Python Programming Essentials for AI-102

You don’t need to be an expert, but you should be comfortable writing scripts and using APIs.

🔤 Core Python Skills

• Variables, data types, loops, functions
• File I/O (especially reading text and images)
• Working with JSON (parsing API responses)

🧪 Libraries to Know

• requests – for REST API calls
• json – for handling JSON responses
• pandas – for data manipulation (basic level)
• matplotlib or plotly – optional, for basic visualizations

🧠 AI/ML Libraries (familiarity helps)

• scikit-learn
• numpy, pillow (for image data)
• transformers (if using NLP from Hugging Face — optional)

---

✅ 4. Basics of Azure Cognitive Services, REST APIs, and SDKs

🌍 REST API Concepts

• HTTP methods: GET, POST, PUT, DELETE
• Status codes: 200, 400, 401, 403, 404, 500
• Headers (especially Ocp-Apim-Subscription-Key)
• Authentication via API keys or Azure AD tokens
• Making calls with tools like Postman or Python requests

🧩 Azure SDKs

• Python SDK examples:
  • azure-cognitiveservices-vision-computervision
  • azure-ai-textanalytics
  • azure-identity
  • azure-search-documents
• Concepts of client objects, credentials, request/response structure

💡 Using the SDK vs. REST API

• SDK is easier for Python devs
• REST API is more universal & helps you understand under-the-hood behavior
• SDK wraps the REST API and handles auth, error handling, etc.