In a world where artificial intelligence is reshaping industries and transforming the way we live, Azure OpenAI stands at the forefront, embodying the boundless possibilities of this futuristic technology. But what does Azure OpenAI truly represent? It is an exciting combination of Microsoft’s cloud infrastructure with OpenAI’s state-of-the-art AI models, creating a fusion of innovation and intelligence. With a single question, Azure OpenAI can delve into vast amounts of data, make predictions, understand human language, and even generate human-like text—all at an excellent scale. Join our journey in Mastering Azure OpenAI Service to uncover the secret of prompt engineering and learn the potential of Azure OpenAI services. Furthermore, as we conclude this article, we will provide you with a quick guide on how to create your own AI Service using OpenAI. This straightforward guide will assist you in setting up your personal AI Service, empowering you to delve into the world of AI with ease and confidence.

What is Azure Open AI?

Azure Open AI Service is a cloud-based service provided by Microsoft Azure that offers access to OpenAI’s advanced artificial intelligence models and technologies. It allows developers to leverage the power of OpenAI’s language models, such as GPT-3, to build intelligent applications, generate human-like text, automate conversations, and perform various natural language processing tasks.

By integrating Azure Open AI Service into their applications, developers can take advantage of cutting-edge AI capabilities to enhance customer experiences, automate workflows, and enable intelligent decision-making. The service provides a range of APIs and tools that simplify the process of integrating AI capabilities into applications, making it easier for developers to harness the power of OpenAI’s models without requiring extensive expertise in machine learning.

Azure Open AI Service offers pre-trained models that cover a wide range of use cases, including chatbots, language translation, content generation, sentiment analysis, and more. These models have been trained on vast amounts of data and can generate high-quality and contextually relevant responses.

Additionally, Azure Open AI Service provides features for prompt engineering, which involves fine-tuning and customizing the behavior of AI models by providing specific instructions or examples. This allows developers to shape the output of the models to suit their specific needs and ensure more accurate and desired results. Azure Open AI Service empowers developers to leverage state-of-the-art AI capabilities, enhance their applications with natural language understanding and generation, and create more intelligent and interactive experiences for their users.

Fundamentals of Azure Open AI

In the world of Mastering Azure OpenAI Service: Prompt Engineering Techniques and More, we will explore the key concepts of prompt engineering, prompt engineering techniques, content filtering, and embeddings. These elements play a crucial role in harnessing the power of Azure Open AI and creating effective and efficient AI solutions. Let’s delve into each of these aspects to understand their significance in building intelligent applications powered by Azure Open AI.

What is Prompt Engineering?

Prompt engineering is the process of designing and refining prompts or instructions provided to the language model to achieve desired outputs or responses. It involves crafting the input text in a way that elicits the desired behavior or generates the intended output from the model.

In the context of Azure OpenAI, prompt engineering is particularly relevant when using models like GPT-3 (Generative Pre-trained Transformer) for tasks such as text generation or completion. GPT-3 is a powerful language model that can generate coherent and contextually relevant text, but it requires well-crafted prompts to produce accurate and useful results.

Effective prompt engineering involves considering the following aspects:

1. Clarity and specificity: Prompts should be clear, concise, and specific about the desired task or question. Vague or ambiguous prompts may lead to unexpected or irrelevant responses from the model.
2. Formatting and context: Providing the necessary context and structure within the prompt is crucial. Including relevant details, instructions, or examples can guide the model to produce more accurate and relevant outputs.
3. System behavior specification: Prompt engineering can be used to guide the model’s behavior. By explicitly instructing the model to follow certain guidelines or policies, developers can influence the generated responses to align with specific requirements or ethical considerations.
4. Control tokens: Azure OpenAI provides control tokens that allow fine-grained control over the behavior of the language model. By using these tokens strategically within prompts, developers can influence attributes like tone, style, or content generation.
5. Iterative refinement: Prompt engineering often involves an iterative process of experimentation and refinement. Developers can test different prompts, analyze the model’s responses, and make adjustments to optimize the desired outputs.

By leveraging prompt engineering techniques in Azure OpenAI, developers can harness the capabilities of powerful language models like GPT-3 to generate high-quality and contextually appropriate text for various applications such as chatbots, content generation, language translation, and more.

Creating successful prompts often involves utilizing a technique known as:

1. One-shot Learning: A technique where the model is provided with a single example of a task or concept, allowing it to generate the result based on that single example.
2. Few-shot Learning: A technique where the model is provided with more than one example for a task, allowing it to generate the result based on those limited examples.

One-shot learning and few-shot learning

One-shot learning and few-shot learning are techniques in machine learning that aim to train models to recognize or classify new objects or concepts with very limited or scarce training examples.

1. One-shot learning: One-shot learning refers to the ability of a model to recognize or classify new instances based on just a single example. Traditional machine learning algorithms often require a large amount of labeled training data to learn patterns and make accurate predictions. However, in one-shot learning, the model is trained to generalize from a single example and make predictions about unseen instances.

For example, Suppose you want to build a model that can identify a specific type of bird, let’s say a “Purple-crowned Fairywren,” which is a rare and unique bird species. In one-shot learning, you would provide the model with just a single image of the Purple-crowned Fairywren. The model learns to recognize the distinct features and patterns present in that image. When a new image of a Purple-crowned Fairywren is presented to the model, it uses the knowledge gained from the single example to make predictions. Despite having only seen one instance, the model can identify whether the new image contains a Purple-crowned Fairywren or not.

2. Few-shot learning: Few-shot learning extends the concept of one-shot learning by allowing the model to learn from a few examples instead of just one. It recognizes that having more than one example can improve the model’s ability to generalize and make accurate predictions. In few-shot learning, the model is trained on a small labeled dataset, typically containing a few examples from each class or category. This limited training set is used to learn the underlying patterns and features that distinguish different classes. The model then applies this knowledge to classify new instances.

For instance, now, let’s consider few-shot learning, where the model is trained with a few examples from each class. Continuing with the bird classification scenario, you have a dataset with five images each of five different bird species: Purple-crowned Fairywren, Blue Jay, Cardinal, Hummingbird, and Woodpecker. The model is trained on this limited dataset, learning the distinguishing features and patterns of each bird species. It generalizes from the small number of examples to identify the unique characteristics of each class. When you present a new image of a bird, say a species not present in the training dataset, the model applies its understanding of different bird species. It can leverage the learned features to make a prediction about the unseen bird’s species based on the similarities it identifies with the known bird classes. In this way, few-shot learning allows the model to handle novel instances by generalizing from a small number of training examples and leveraging its learned knowledge.

Difference between One-shot learning vs Few-Shot Learning

Both one-shot learning and few-shot learning are important techniques in scenarios where obtaining a large labeled dataset for training is challenging or impractical. They enable models to learn from limited examples and generalize well to new, unseen instances. These techniques have applications in areas such as image recognition, object detection, natural language processing, and computer vision, where data scarcity is a common challenge.

One-shot learning and few-shot learning are both significant, yet they differ in terms of complexity, performance, and usage. Let’s explore the dissimilarities between One-shot learning and Few-shot Learning:

Prompt Engineering Techniques

The principles of prompt engineering can be generalized with different types of models. There are basically two APIs where prompt engineering comes and each API requires input data to be formatted differently which impacts the whole prompt design:

1. Chat Completion API: Supports ChatGPT and GPT-4 models, which take the input formatted in a specific chat-like transcript stored inside a dictionary.
2. Completion API: Supports old GPT-3 model and is much more flexible input requirements. This takes the text input with no specific format.

While using prompt engineering you need to validate the responses which are generated by the model. To validate the responses, here are some of the prompt engineering techniques which will help you to increase the accuracy and grounding of responses you generate with a Large Language Model (LLM):

Technique 1: Importance of System Message

A system message refers to a specific type of message or instruction provided at the beginning of a conversation or prompts in a conversational AI system. It is a message intended to guide and set the context for the AI model or assistant’s behavior and responses. Here’s a simplified explanation:

Imagine a conversation between a user and an AI assistant. The system message is the initial message that the assistant presents to the user before any user input is received. It is like an introduction or a set of instructions provided by the AI system.

The purpose of a system message is to prime the AI model with important information, instructions, or context. It helps define the assistant’s personality, behavior, or limitations, and can also specify the format or expected responses. The system message sets the stage for the conversation and provides initial guidance to the assistant.

For example, a system message might say, “Hello! I’m an AI assistant designed to help with travel information. Feel free to ask me about flights, hotels, or attractions. Please note that I may not have real-time data, so always double-check for the latest information.”

In this example, the system message informs the user about the assistant’s role and limitations. It sets the expectation that the assistant can assist with travel-related queries but may not have real-time information. This helps the user understand how to interact with the assistant and manage their expectations.

System messages are valuable for providing important context and instructions to the AI model or assistant, enabling more effective and accurate responses. They allow the AI system to align its behavior with the desired functionality and enhance the user experience by providing clarity and guidance from the beginning of the conversation.

Technique 2: Use Few-Shot Learning

Few-shot learning is a machine learning approach that aims to train models to learn new tasks or recognize new concepts with only a limited amount of labeled data. In traditional machine learning, models often require large amounts of labeled data to achieve good performance. However, in few-shot learning, the goal is to train models that can generalize and adapt to new tasks or classes quickly, even when provided with only a few examples.

The idea behind few-shot learning is to leverage prior knowledge learned from related tasks to perform well on new, unseen tasks. Instead of training a model from scratch for each new task, few-shot learning algorithms learn to extract generalizable knowledge that can be fine-tuned or adapted to specific tasks using only a few examples.

For example, let’s say you want to train a model to recognize different breeds of dogs, but you only have a small number of labeled images for each breed. In a few-shot learning scenario, you would provide the model with a few examples of each breed during the training phase. The model would then learn to generalize from this limited labeled data and be able to recognize new dog breeds it hasn’t seen before.

Few-shot learning has gained attention in the field of artificial intelligence as it reduces the reliance on vast amounts of labeled data and enables models to adapt quickly to new tasks. It has applications in various domains, including computer vision, natural language processing, and reinforcement learning.

Technique 3: Use Chat Completion API for Conversations and Non-Chat Scenarios

The Chat Completion API is a tool provided by OpenAI that enables you to use language models for generating responses in conversational scenarios. It is designed to handle back-and-forth interactions, making it suitable for chat-based applications and dialogue systems.

When using the Chat Completion API for conversations, you can provide a series of messages as the conversation history. Each message includes a role (such as “system”, “user”, or “assistant”) and the corresponding content. The conversation history helps provide context for the model, allowing it to generate more relevant responses based on the previous messages.

For example, let’s say you have a chat-based customer support application. You can send a conversation history like this:

[  {"role": "system", "content": "You are an AI customer support assistant."},  
   {"role": "user", "content": "I have a problem with my order."},  
   {"role": "assistant", "content": "I'm sorry to hear that. Could you please provide me with your 
   order number?"}]

By using the Chat Completion API with this conversation history, you can generate the next response from the assistant based on the user’s input and the context of the conversation.

In addition, to chat scenarios, the Chat Completion API can also be used in non-chat scenarios where you have a single-turn task or prompt. In this case, you can provide a single message as the prompt without any conversation history. The model will generate a response based on the provided prompt.

For example, in a non-chat scenario, you can use the API like this:

[  {"role": "user", "content": "Translate the following English text to French: 'Hello, how are you?'"}]

The API will generate the translated French text based on the prompt.

The Chat Completion API is versatile and can be used for both interactive conversations and single-turn prompts, depending on your application’s requirements. It allows you to harness the power of language models for a wide range of chat-based and non-chat scenarios, enabling more dynamic and engaging interactions with users.

Technique 4: You should Start your prompt with Clear Instructions

Starting your prompt with clear instructions means providing explicit and unambiguous guidance to the AI model right at the beginning of your input. It helps set the expectations and informs the model about the desired behavior and the task it should perform.

Clear instructions can include specific details about what you want the model to do, the format you want the response in or any limitations or constraints you want the model to follow. By stating these instructions upfront, you provide guidance to the AI model on how to generate a relevant and accurate response.

For example, let’s say you want the AI model to write a short story about a magical adventure. Instead of starting with an open-ended question like, “Tell me a story,” you can provide clear instructions like, “Please write a short story about a young wizard who discovers a hidden treasure in a mystical forest.”

By starting with clear instructions, you guide the model to focus on a specific theme or task, ensuring that the generated response aligns with your desired outcome.

Clear instructions are essential to ensure that the AI model understands the task, avoids confusion, and produces more relevant and coherent responses. It helps to establish a clear communication channel between you and the AI model, enabling you to get the desired results more effectively.

Technique 5: Repeat the Instructions at the end of the prompt

Repeating the instructions at the end of the prompt means restating the guidelines or directions given to the AI model after providing the conversation history or system message. It helps reinforce the instructions and ensures that the model understands and follows them throughout the conversation.

For example, let’s say we have an AI model that helps with math problems. The instructions provided in the prompt could be: “Hi, I’m the Math Assistant! Please provide me with a math problem, and I’ll do my best to solve it for you.”

To emphasize these instructions, we can repeat them at the end of the prompt: “5 + 3 = ?” “Hi, I’m the Math Assistant! Please provide me with a math problem, and I’ll do my best to solve it for you.”

Repeating the instructions reminds the AI model about its purpose and helps maintain consistency in its responses. It ensures that the model understands its role and knows what kind of input it should expect from the user.

Repeating instructions at the end of the prompt is a helpful technique to ensure clarity and reinforce the expected behavior of the AI model.

Technique 6: Add more instruction at the end of the prompt

Adding more instructions at the end of a prompt refers to providing additional guidance, clarification, or specific requirements to the AI model or assistant after the initial instructions. This technique aims to further guide the model’s behavior and ensure a desired output. Here’s a simplified explanation:

When you include more instructions at the end of a prompt, you are giving extra information or direction to the AI model to help it understand and fulfill the task more effectively. These additional instructions can include specific details, constraints, examples, or reminders related to the task at hand.

The purpose of including more instructions at the end is to provide further clarity, address potential ambiguities, or highlight important aspects that the model should consider while generating a response. By adding these supplementary instructions, you can guide the model’s behavior toward the desired outcome and ensure it aligns with your specific requirements.

For example, if the initial prompt instructs the model to write a story about a magical adventure, the additional instructions at the end could specify that the story should include a talking animal companion, a hidden treasure, and a happy ending. These extra instructions give the model more guidance on what elements to include in the story and help ensure that it meets the desired criteria.

By incorporating more instructions at the end of the prompt, you can provide the AI model with clearer guidance, improve its understanding of the task, and enhance the likelihood of generating responses that align with your expectations.

Technique 7: Write clear and organized Syntax

Writing clear and organized syntax refers to structuring the prompt or instructions in a well-structured and easy-to-understand manner. It involves using proper grammar, formatting, and logical sequencing of the information to ensure clarity and readability. Here’s a simplified explanation:

When you write clear and organized syntax, you are presenting the instructions or prompts in a way that is easy to follow and comprehend. This includes using correct grammar, punctuation, and sentence structure to convey your message effectively. It also involves organizing the information in a logical and sequential manner, so it flows naturally and is easy to interpret.

The purpose of using clear and organized syntax is to ensure that the AI model or assistant can understand the instructions accurately. By presenting the information in a clear and coherent manner, you minimize the chances of misinterpretation or confusion. This allows the model to generate responses that align with your intended meaning and fulfill the desired task.

For example, when providing a set of instructions for a task, you would use clear and organized syntax by using bullet points or numbered lists to break down the steps. Each step would be written as a clear and concise statement, making it easy for the model to understand and follow the sequence of actions.

By writing clear and organized syntax, you improve the readability and comprehension of the instructions for the AI model. This enhances the model’s ability to interpret and execute the task accurately, leading to more coherent and relevant responses.

Technique 8: Break down the task

When you break down a task, you analyze it carefully and identify the individual components or steps required to complete it successfully. Instead of tackling the entire task at once, you divide it into smaller subtasks that can be addressed one by one.

The purpose of breaking down the task is to make it more understandable and approachable for the AI model. By breaking it into smaller parts, you provide the model with clear instructions and a step-by-step approach to follow. This helps prevent confusion or overwhelm that may arise when dealing with a complex task as a whole.

For example, if the task is to write a research paper, you would break it down into subtasks such as conducting research, outlining the paper, writing the introduction, developing the main arguments, and concluding the paper. By breaking down the task in this way, you provide the model with clear and sequential instructions on how to proceed.

Breaking down the task helps the AI model to understand the specific requirements of each step and focus on accomplishing them individually. It allows for a more systematic and structured approach, making the overall task more manageable and increasing the likelihood of generating accurate and coherent responses.

Technique 9: Use Affordances

Using affordances refers to providing cues, hints, or contextual information in the prompt that guide the AI model or assistant toward the desired behavior or output. Affordances are designed to make it easier for the model to understand the task or prompt by highlighting relevant information or indicating the expected actions. Here’s a simplified explanation:

When you use affordances, you include additional information or clues in the prompt to help the AI model better understand what is expected of it. These cues can take various forms, such as examples, explicit instructions, or contextual details that guide the model’s behavior.

The purpose of using affordances is to provide the AI model with additional guidance or context that aids its decision-making process. Affordances help the model recognize patterns, make connections, and generate responses that align with the desired output.

For example, if the task is to generate a creative story, an affordance could be providing a starting sentence or a brief outline to give the model a direction to follow. This helps the model understand the structure and style expected for the story and can serve as a prompt to trigger its creativity.

By incorporating affordances, you make it easier for the AI model to understand and generate responses that meet your expectations. Affordances provide clarity and guidance, reducing ambiguity and increasing the chances of the model producing more relevant and accurate outputs. It enhances the model’s understanding by providing helpful cues and hints, resulting in more effective and tailored responses.

Technique 10: Provide Grounding Context

Providing grounding context refers to including relevant background information or context in the prompt or conversation to help the AI model or assistant understand the specific scenario, topic, or context in which it is expected to generate responses. Grounding context provides a frame of reference for the model, aiding its comprehension and enabling more accurate and relevant outputs. Here’s a simplified explanation:

When you provide grounding context, you offer additional details or background information to the AI model that help it understand the specific situation or subject matter. This context can include relevant facts, historical background, specific constraints, or any other information that is crucial for the model to generate appropriate responses.

For example, in a conversation with an AI assistant about booking a hotel room, providing grounding context could involve specifying the desired location, dates of stay, budget constraints, and any specific requirements or preferences. This information helps the assistant provide more tailored recommendations or options that meet the user’s needs.

By offering grounding context, you enable the AI model to generate responses that are more informed and relevant to the given situation. It helps the model consider the specific circumstances and constraints, resulting in more personalized and contextually appropriate outputs.

Content Filtering

Content filtering refers to the process of screening and controlling the type of information or content that is allowed or restricted within a particular system, platform, or environment. It involves implementing mechanisms or algorithms to filter out undesirable or inappropriate content based on predefined rules or criteria. Here’s a simplified explanation:

Content filtering is like having a filter or sieve that checks and sorts information to determine what is acceptable and what should be blocked or restricted. It is used to ensure that the content available in a system or platform aligns with certain standards or guidelines.

The purpose of content filtering is to regulate the access to and dissemination of content to maintain safety, security, and compliance within a specific context. It can be used to prevent the exposure of inappropriate, harmful, or sensitive material, such as explicit or violent content, spam, or fraudulent information.

For example, content filtering can be implemented in a social media platform to block or flag posts that contain hate speech, nudity, or other forms of harmful content. It can also be used in educational institutions or workplaces to restrict access to certain websites or online resources that are deemed inappropriate or irrelevant to the intended purpose.

Content filtering can be done through various techniques, including keyword-based filtering, image recognition, URL blacklisting, or machine learning algorithms that analyze content based on patterns or predefined categories.

By implementing content filtering mechanisms, organizations and platforms can create a safer and more controlled environment for their users. It helps ensure compliance with legal regulations, community guidelines, or specific user requirements, promoting a positive and secure user experience.

Embedding

Think of embedding as a way to convert words or sentences into numbers. It’s like assigning a unique code or address to each word or piece of text. This numerical representation captures the meaning or context of the text in a way that machines can understand.

The purpose of embedding is to capture the semantic relationships and similarities between words or sentences. By representing them as vectors in a high-dimensional space, words with similar meanings or contexts tend to have similar vector representations. This allows mathematical models to perform computations and make predictions based on the relationships between these vectors.

For example, in a word embedding model, the word “king” might be represented by a vector like [0.2, 0.4, -0.1], while the word “queen” could be represented by [0.15, 0.35, -0.05]. The model learns to assign these vectors based on the co-occurrence patterns of words in a large corpus of text. The similarity between the words “king” and “queen” can be computed by measuring the distance or angle between their respective vectors.

Embedding is widely used in various NLP tasks, such as text classification, sentiment analysis, machine translation, and question-answering. It allows machines to understand and process textual data by mapping it to a numerical representation that encodes important linguistic properties.

By using embedding techniques, we can leverage the power of machine learning algorithms to analyze and understand text data. Embeddings help capture the nuances and relationships within language, enabling models to perform tasks like understanding meaning, making predictions, or generating relevant responses.

How to Quick Start with Azure Open AI?

To get started with Azure OpenAI, you first need to deploy the cognitive services. This involves creating and configuring the necessary resources in Azure. The Cognitive Services provide a range of pre-built AI capabilities that can be easily integrated into your applications.

Follow the below steps to deploy the cognitive service to create Azure OpenAI:

STEP 1: Open Azure Portal and click on “Create a resource“.

STEP 2: Now, Select “Cognitive Services” in Azure Marketplace.

STEP 3: Click on “Create” to begin the process.

STEP 4: Enter the details such as subscription, Region where you want to deploy the cognitive service and Name of Azure OpenAI Service. Click Next.

Till now, only the standard tier is available. In the future, if other pricing tiers are available, you can select accordingly.

STEP 5: Now, select whether you want to configure all network security for cognitive services or disable it from all networks and make it only accessible from private endpoints. Click Next.

STEP 6: Enter the name and value for the cognitive service which will categorize resources and view consolidating billing with the same tag to multiple resources and resource groups.

STEP 7: Now, review all the details which you have entered till now and click on the “Create” button.

Azure OpenAI is created.

STEP 8: Now, go to Resource Management => Keys and Endpoint. In the right panel, click on “Show Keys“. You will see the Endpoint URL. See the below image to better understand.

Now you have to copy the URL because this URL will be used in the open.api_key.

Once you finish setting up the Cognitive Services, your instance will be ready to use. Now, you can take advantage of Azure OpenAI by combining Cognitive Services with your own applications. In this example, I am using Databricks Notebook to access the cognitive service by service endpoint. But you can use simple Python code to achieve the same.

Below, we have a proper code example:

%pip install --upgrade openai

import os
import openai
openai.api_type = "azure"
openai.api_base = "This is the Endpoint URL taken from step 8"
openai.api_version = "2022-12-01"
openai.api_key = "When you press SHOW key, it will provide the key which you need to put here"

response = openai.Completion.create(
  engine="text-davinci-003",
  prompt="Write a job description for the following job title: 'Business Intelligence Analyst'. The job description should outline the main responsibilities of the role, list the required qualifications, highlight unique benefits like flexible working hours, and provide information on how to apply.\n\nBusiness Intelligence",
  temperature=1,
  max_tokens=600,
  top_p=1,
  frequency_penalty=0,
  presence_penalty=0,
  best_of=1,
  stop=None)

The above code will install and upgrade the OpenAI library using the %pip command. It then imports the necessary modules and sets up the OpenAI API with specific configurations.

After that, it makes a request to the OpenAI API to create a completion. The completion request is made using the openai.Completion.create() method. The method takes various parameters to customize the completion, such as the engine to use, the prompt text, temperature, max tokens, and other settings.

In this specific example, the prompt text is a job description request for the position of “Business Intelligence Analyst.” The code sends this prompt to the OpenAI API and expects a response with completion that describes the job responsibilities, qualifications, benefits, and application instructions.

Once the completion response is received, it can be further processed or used in any way necessary in the rest of the code.

Here is the output of the Open AI:

Analyst:

We are looking for an experienced and highly motivated Business Intelligence Analyst to join our company. The successful candidate should have a knowledge and skill set in the analysis and interpretation of data using various techniques and tools.

Responsibilities:

• Gathering, interpreting, and modeling data to identify trends and helpful insights
• Developing and maintaining important data sources and data warehouses for use in reporting and analysis
• Performing analysis to identify opportunities to improve business practices
• Preparing and presenting valuable insights to internal stakeholders
• Producing accurate forecasts and reports to support business decisions
• Updating and maintaining ETL processes Qualifications:
• Bachelor’s degree in Data Science, Statistics, Computer Science, or a related field
• Experience with Business Intelligence technology such as Tableau and Qlik
• Proficiency with SQL

Conclusion

This blog focuses on Azure Open AI and its fundamentals, including prompt engineering, one-shot learning, few-shot learning, and various prompt engineering techniques. We have also delved into content filtering and embedding, along with a quick start guide. Azure Open AI enables developers to build robust and intelligent AI applications. Through prompt engineering and leveraging Azure Open AI’s capabilities, developers can enhance AI model performance and develop contextually aware experiences.