Azure AI Services Integration Guide

Azure AI Overview

Microsoft Azure provides a comprehensive suite of AI services:

• Azure OpenAI Service - GPT models with enterprise features
• Azure Cognitive Services - Vision, Speech, Language, Decision
• Azure Machine Learning - Custom model training and deployment
• Azure AI Search - Intelligent search with semantic ranking

Getting Started with Azure

Prerequisites

Azure subscription (free tier available)

Azure CLI installed

Basic understanding of REST APIs

Creating Resources

# Login to Azure
az login

# Create resource group
az group create --name myai-rg --location eastus

# Create Azure OpenAI resource
az cognitiveservices account create \
  --name myopenai-resource \
  --resource-group myai-rg \
  --kind OpenAI \
  --sku S0 \
  --location eastus

Azure OpenAI Service

Why Azure OpenAI?

• Enterprise security and compliance
• Virtual network support
• Content filtering
• Regional deployment options
• SLA guarantees

Deployment

# Deploy a model
az cognitiveservices account deployment create \
  --name myopenai-resource \
  --resource-group myai-rg \
  --deployment-name gpt-4-deployment \
  --model-name gpt-4 \
  --model-version "0613" \
  --model-format OpenAI \
  --sku-capacity 10 \
  --sku-name Standard

Using the API

from openai import AzureOpenAI

client = AzureOpenAI(
    api_key=os.environ["AZURE_OPENAI_KEY"],
    api_version="2024-02-15-preview",
    azure_endpoint=os.environ["AZURE_OPENAI_ENDPOINT"]
)

response = client.chat.completions.create(
    model="gpt-4-deployment",  # Your deployment name
    messages=[
        {"role": "system", "content": "You are a helpful assistant."},
        {"role": "user", "content": "Hello!"}
    ]
)

print(response.choices[0].message.content)

Streaming

stream = client.chat.completions.create(
    model="gpt-4-deployment",
    messages=[{"role": "user", "content": "Write a story"}],
    stream=True
)

for chunk in stream:
    if chunk.choices[0].delta.content:
        print(chunk.choices[0].delta.content, end="")

Azure Cognitive Services

Computer Vision

Analyze images for objects, text, and more:

from azure.cognitiveservices.vision.computervision import ComputerVisionClient
from msrest.authentication import CognitiveServicesCredentials

client = ComputerVisionClient(
    endpoint=os.environ["VISION_ENDPOINT"],
    credentials=CognitiveServicesCredentials(os.environ["VISION_KEY"])
)

# Analyze an image
analysis = client.analyze_image(
    url="https://example.com/image.jpg",
    visual_features=["Categories", "Description", "Tags", "Objects"]
)

print(f"Description: {analysis.description.captions[0].text}")
for tag in analysis.tags:
    print(f"Tag: {tag.name} ({tag.confidence:.2%})")

Speech Services

Text-to-speech and speech-to-text:

import azure.cognitiveservices.speech as speechsdk

# Speech to Text
speech_config = speechsdk.SpeechConfig(
    subscription=os.environ["SPEECH_KEY"],
    region=os.environ["SPEECH_REGION"]
)

audio_config = speechsdk.AudioConfig(filename="audio.wav")
recognizer = speechsdk.SpeechRecognizer(
    speech_config=speech_config,
    audio_config=audio_config
)

result = recognizer.recognize_once()
print(f"Recognized: {result.text}")

# Text to Speech
synthesizer = speechsdk.SpeechSynthesizer(speech_config=speech_config)
result = synthesizer.speak_text_async("Hello, world!").get()

Language Services

Text analytics, translation, and more:

from azure.ai.textanalytics import TextAnalyticsClient
from azure.core.credentials import AzureKeyCredential

client = TextAnalyticsClient(
    endpoint=os.environ["LANGUAGE_ENDPOINT"],
    credential=AzureKeyCredential(os.environ["LANGUAGE_KEY"])
)

# Sentiment analysis
documents = ["I love this product!", "This is terrible."]
response = client.analyze_sentiment(documents)

for doc in response:
    print(f"Sentiment: {doc.sentiment}")
    print(f"Scores: positive={doc.confidence_scores.positive:.2%}, "
          f"negative={doc.confidence_scores.negative:.2%}")

Translator

import requests

endpoint = "https://api.cognitive.microsofttranslator.com"
path = "/translate?api-version=3.0&to=es&to=fr"

headers = {
    "Ocp-Apim-Subscription-Key": os.environ["TRANSLATOR_KEY"],
    "Ocp-Apim-Subscription-Region": os.environ["TRANSLATOR_REGION"],
    "Content-Type": "application/json"
}

body = [{"text": "Hello, world!"}]
response = requests.post(endpoint + path, headers=headers, json=body)

for translation in response.json()[0]["translations"]:
    print(f"{translation['to']}: {translation['text']}")

Azure AI Search

Intelligent search with semantic ranking:

from azure.search.documents import SearchClient
from azure.core.credentials import AzureKeyCredential

client = SearchClient(
    endpoint=os.environ["SEARCH_ENDPOINT"],
    index_name="my-index",
    credential=AzureKeyCredential(os.environ["SEARCH_KEY"])
)

# Semantic search
results = client.search(
    search_text="What is machine learning?",
    query_type="semantic",
    semantic_configuration_name="my-semantic-config",
    select=["title", "content"],
    top=5
)

for result in results:
    print(f"Title: {result['title']}")
    print(f"Score: {result['@search.score']}")

RAG with Azure

Retrieval-Augmented Generation combining Azure AI Search and Azure OpenAI:

# 1. Search for relevant documents
search_results = search_client.search(
    search_text=user_query,
    query_type="semantic",
    top=5
)

# 2. Build context from results
context = "\n".join([doc["content"] for doc in search_results])

# 3. Generate response with context
response = openai_client.chat.completions.create(
    model="gpt-4-deployment",
    messages=[
        {
            "role": "system",
            "content": f"Answer based on this context:\n{context}"
        },
        {"role": "user", "content": user_query}
    ]
)

Best Practices

Security

• Use managed identities where possible
• Store keys in Azure Key Vault
• Enable private endpoints for sensitive workloads
• Implement content filtering

Cost Management

• Use appropriate model sizes
• Implement caching for repeated queries
• Monitor usage with Azure Cost Management
• Consider provisioned throughput for predictable workloads

Performance

• Deploy to regions close to users
• Use streaming for better UX
• Implement retry logic with exponential backoff
• Cache embeddings for frequently accessed content

Error Handling

from azure.core.exceptions import HttpResponseError

try:
    response = client.chat.completions.create(...)
except HttpResponseError as e:
    if e.status_code == 429:
        print("Rate limited - implement backoff")
    elif e.status_code == 401:
        print("Authentication failed - check credentials")
    else:
        print(f"Error: {e.message}")

Conclusion

Azure AI Services provide enterprise-grade AI capabilities with the security and compliance features organizations need. Whether using Azure OpenAI for language models or Cognitive Services for specialized tasks, the platform offers comprehensive tools for building AI applications.

---

Next: Google Vertex AI Integration - Working with Google's AI platform