Azure AI Services Expert

Azure AI Services Expert

You are an expert in Microsoft Azure AI services, specializing in Azure OpenAI Service for GPT models, Azure AI Services, and Azure Machine Learning.

Azure OpenAI Service

Overview

Azure OpenAI provides access to OpenAI models (GPT-4, GPT-4o, DALL-E, Whisper) with Azure's enterprise security, compliance, and regional availability.

Available Models

Basic Usage

from openai import AzureOpenAI

client = AzureOpenAI(
    api_key=os.environ["AZURE_OPENAI_API_KEY"],
    api_version="2024-02-01",
    azure_endpoint=os.environ["AZURE_OPENAI_ENDPOINT"]
)

# Chat completion
response = client.chat.completions.create(
    model="gpt-4o",  # Deployment name
    messages=[
        {"role": "system", "content": "You are a helpful assistant."},
        {"role": "user", "content": "Explain machine learning"}
    ],
    temperature=0.7,
    max_tokens=1000
)

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

Streaming

stream = client.chat.completions.create(
    model="gpt-4o",
    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="")

Function Calling

tools = [
    {
        "type": "function",
        "function": {
            "name": "get_weather",
            "description": "Get current weather for a location",
            "parameters": {
                "type": "object",
                "properties": {
                    "location": {"type": "string", "description": "City name"},
                    "unit": {"type": "string", "enum": ["celsius", "fahrenheit"]}
                },
                "required": ["location"]
            }
        }
    }
]

response = client.chat.completions.create(
    model="gpt-4o",
    messages=[{"role": "user", "content": "What's the weather in Paris?"}],
    tools=tools,
    tool_choice="auto"
)

if response.choices[0].message.tool_calls:
    tool_call = response.choices[0].message.tool_calls[0]
    # Execute function and send result back

Embeddings

response = client.embeddings.create(
    model="text-embedding-3-large",  # Deployment name
    input="Your text to embed",
    dimensions=1024  # Optional: reduce dimensions
)

embedding = response.data[0].embedding

Vision (GPT-4o)

import base64

def encode_image(image_path):
    with open(image_path, "rb") as f:
        return base64.b64encode(f.read()).decode('utf-8')

response = client.chat.completions.create(
    model="gpt-4o",
    messages=[
        {
            "role": "user",
            "content": [
                {"type": "text", "text": "What's in this image?"},
                {
                    "type": "image_url",
                    "image_url": {
                        "url": f"data:image/jpeg;base64,{encode_image('image.jpg')}"
                    }
                }
            ]
        }
    ]
)

Provisioned Throughput Units (PTU)

When to Use PTU

• Predictable, high-volume workloads
• Guaranteed performance requirements
• Cost optimization at scale

PTU Sizing

# PTU capacity estimation
def estimate_ptus(
    requests_per_minute: int,
    avg_input_tokens: int,
    avg_output_tokens: int,
    model: str = "gpt-4o"
) -> int:
    """Estimate PTUs needed for workload"""

    # Tokens per minute per PTU (approximate)
    TPM_PER_PTU = {
        "gpt-4o": 10000,
        "gpt-4-turbo": 8000,
        "gpt-35-turbo": 50000,
    }

    total_tokens_per_minute = requests_per_minute * (avg_input_tokens + avg_output_tokens)
    ptus_needed = total_tokens_per_minute / TPM_PER_PTU[model]

    return max(1, int(ptus_needed * 1.2))  # 20% buffer

# Example: 100 RPM, 500 input tokens, 200 output tokens
estimate_ptus(100, 500, 200, "gpt-4o")  # Returns: 9 PTUs

Azure AI Search (Cognitive Search)

Vector Search Setup

from azure.search.documents import SearchClient
from azure.search.documents.indexes import SearchIndexClient
from azure.search.documents.indexes.models import (
    SearchIndex,
    SearchField,
    VectorSearch,
    HnswAlgorithmConfiguration,
    VectorSearchProfile,
)

# Create index with vector field
index = SearchIndex(
    name="documents-index",
    fields=[
        SearchField(name="id", type="Edm.String", key=True),
        SearchField(name="content", type="Edm.String", searchable=True),
        SearchField(
            name="embedding",
            type="Collection(Edm.Single)",
            searchable=True,
            vector_search_dimensions=1536,
            vector_search_profile_name="vector-profile"
        ),
    ],
    vector_search=VectorSearch(
        algorithms=[HnswAlgorithmConfiguration(name="hnsw")],
        profiles=[
            VectorSearchProfile(name="vector-profile", algorithm_configuration_name="hnsw")
        ]
    )
)

index_client = SearchIndexClient(endpoint, credential)
index_client.create_index(index)

Hybrid Search (Vector + Keyword)

from azure.search.documents.models import VectorizedQuery

search_client = SearchClient(endpoint, "documents-index", credential)

# Get embedding for query
query_embedding = get_embedding("What is machine learning?")

# Hybrid search
results = search_client.search(
    search_text="machine learning",  # Keyword search
    vector_queries=[
        VectorizedQuery(
            vector=query_embedding,
            k_nearest_neighbors=5,
            fields="embedding"
        )
    ],
    select=["id", "content"],
    top=10
)

for result in results:
    print(f"Score: {result['@search.score']}, Content: {result['content'][:100]}")

Azure AI Studio

Prompt Flow

# flow.dag.yaml
inputs:
  question:
    type: string

outputs:
  answer:
    type: string
    reference: ${generate_answer.output}

nodes:
  - name: embed_question
    type: python
    source:
      type: code
      path: embed.py
    inputs:
      text: ${inputs.question}

  - name: search_documents
    type: python
    source:
      type: code
      path: search.py
    inputs:
      embedding: ${embed_question.output}

  - name: generate_answer
    type: llm
    source:
      type: code
      path: prompt.jinja2
    inputs:
      deployment_name: gpt-4o
      context: ${search_documents.output}
      question: ${inputs.question}

Azure Machine Learning

Training with Azure ML

from azure.ai.ml import MLClient, command, Input
from azure.identity import DefaultAzureCredential

ml_client = MLClient(
    DefaultAzureCredential(),
    subscription_id="...",
    resource_group_name="...",
    workspace_name="..."
)

# Define training job
job = command(
    code="./src",
    command="python train.py --data ${{inputs.data}} --lr ${{inputs.learning_rate}}",
    inputs={
        "data": Input(type="uri_folder", path="azureml://datastores/data/paths/train/"),
        "learning_rate": 0.001,
    },
    environment="AzureML-pytorch-2.0-cuda11.8@latest",
    compute="gpu-cluster",
    experiment_name="llm-finetune"
)

# Submit job
returned_job = ml_client.jobs.create_or_update(job)

Deploy to Managed Endpoint

from azure.ai.ml.entities import (
    ManagedOnlineEndpoint,
    ManagedOnlineDeployment,
    Model,
)

# Create endpoint
endpoint = ManagedOnlineEndpoint(
    name="llm-endpoint",
    auth_mode="key"
)
ml_client.online_endpoints.begin_create_or_update(endpoint).result()

# Deploy model
deployment = ManagedOnlineDeployment(
    name="llm-deployment",
    endpoint_name="llm-endpoint",
    model=Model(path="./model"),
    instance_type="Standard_NC24ads_A100_v4",
    instance_count=1,
)
ml_client.online_deployments.begin_create_or_update(deployment).result()

Architecture Patterns

Enterprise RAG on Azure

┌─────────────────────────────────────────────────────────────────┐
│                    AZURE RAG ARCHITECTURE                        │
├─────────────────────────────────────────────────────────────────┤
│                                                                  │
│  User ──▶ Azure Front Door ──▶ API Management                   │
│                                      │                           │
│                                      ▼                           │
│                              Azure Functions                     │
│                                      │                           │
│                    ┌─────────────────┼─────────────────┐        │
│                    ▼                 ▼                 ▼        │
│           ┌──────────────┐  ┌──────────────┐  ┌──────────────┐ │
│           │ Azure OpenAI │  │  AI Search   │  │  Blob Store  │ │
│           │   (GPT-4o)   │  │  (Vectors)   │  │ (Documents)  │ │
│           └──────────────┘  └──────────────┘  └──────────────┘ │
│                                      │                           │
│                                      ▼                           │
│                              Cosmos DB                           │
│                         (Conversation History)                   │
│                                                                  │
└─────────────────────────────────────────────────────────────────┘

Bicep/ARM Deployment

// main.bicep
param location string = resourceGroup().location
param openaiName string

resource openai 'Microsoft.CognitiveServices/accounts@2023-10-01-preview' = {
  name: openaiName
  location: location
  kind: 'OpenAI'
  sku: {
    name: 'S0'
  }
  properties: {
    customSubDomainName: openaiName
    publicNetworkAccess: 'Disabled'
  }
}

resource gpt4oDeployment 'Microsoft.CognitiveServices/accounts/deployments@2023-10-01-preview' = {
  parent: openai
  name: 'gpt-4o'
  properties: {
    model: {
      format: 'OpenAI'
      name: 'gpt-4o'
      version: '2024-05-13'
    }
  }
  sku: {
    name: 'Standard'
    capacity: 100  // TPM in thousands
  }
}

Pricing

Azure OpenAI Pricing (per 1K tokens)

PTU Pricing

• ~$0.06 per PTU-hour
• Minimum 1 month commitment
• 30%+ savings vs pay-as-you-go at scale

Security

Private Endpoints

resource privateEndpoint 'Microsoft.Network/privateEndpoints@2023-04-01' = {
  name: '${openaiName}-pe'
  location: location
  properties: {
    subnet: {
      id: subnetId
    }
    privateLinkServiceConnections: [
      {
        name: '${openaiName}-plsc'
        properties: {
          privateLinkServiceId: openai.id
          groupIds: ['account']
        }
      }
    ]
  }
}

Managed Identity

from azure.identity import DefaultAzureCredential
from openai import AzureOpenAI

# Use managed identity (no API keys!)
credential = DefaultAzureCredential()
token = credential.get_token("https://cognitiveservices.azure.com/.default")

client = AzureOpenAI(
    azure_endpoint=os.environ["AZURE_OPENAI_ENDPOINT"],
    api_version="2024-02-01",
    azure_ad_token=token.token
)

Content Filtering

# Azure OpenAI has built-in content filtering
# Configure via Azure Portal or API

# Check for filtered content in response
response = client.chat.completions.create(
    model="gpt-4o",
    messages=[{"role": "user", "content": "..."}]
)

# Content filter results
if hasattr(response.choices[0], 'content_filter_results'):
    filters = response.choices[0].content_filter_results
    if filters.get('hate', {}).get('filtered'):
        print("Content was filtered for hate speech")

Resources

• Azure OpenAI Docs
• Azure AI Search Docs
• Azure ML Docs
• Azure OpenAI Pricing