Generative AI for Business with Microsoft Azure OpenAI

Outcome: Learn the fundamentals of prompt engineering across text, speech, and image modalities. Gain techniques to design effective prompts for precise AI model outputs across different Azure AI services.

Topics Covered:

Topics Covered:

Coding Track

• Text in Python: Strings and common methods
• File handling: Reading/writing files (.txt, .csv, .json)
• Text manipulation & cleaning: Advanced string operations (find/replace, extract, split/join), handling unwanted characters, dates, and numbers

No-Code Track

Outcome: Apply LLMs to business intelligence and automation tasks, including text-to-text workflows (summarization, translation, content generation) and text-to-label tasks (sentiment analysis, entity recognition).

TechGizmo, a leading tech company, needed to analyze thousands of customer reviews across multiple platforms to identify which product features—such as performance, design, and usability—resonated with users, going beyond basic sentiment analysis.

Outcome: Learn to enhance AI solutions using embeddings, vector databases, and LLMs to build efficient RAG systems for effective question answering.

Topics Covered:

• Business use cases for RAG
• Building RAG workflows with LLMs
• Understanding embeddings and model selection
• Overview of vector databases
• Managing vector databases (CRUD operations)
• Implementing RAG for question answering

Outcome: Implement RAG with multimodal inputs, evaluate outputs, and master advanced techniques to improve the quality and performance of RAG systems in business applications.
Topics Covered:

• RAG with multimodal data sources (text, images, tables)
• Evaluating outputs with Ragas and the RAG Triad
• Advanced retrieval techniques: Hypothetical questions, query expansion
• Advanced generation techniques: Reranking, information compression

Outcome: Learn to deploy and serve LLMs for real-world applications. Gain expertise in prompt engineering, interface design with Streamlit, containerization with Docker, and hosting models.

Topics Covered:

• Deployment and management of LLMs
• Model configuration and optimization for high-performance tasks
• Designing user interfaces with Streamlit
• Containerization with Docker for deployment

Outcome: Understand the evolution, use cases, and architecture of Generative AI agents. Learn agentic design patterns, single-agent systems, and gain hands-on experience building autonomous AI agents.

Topics Covered:

• Generative AI agents: Definition, use cases, LLM vs. agent workflows
• LLM workflows with LangGraph/CrewAI: Sequential, parallel, conditional workflows
• AI agent architecture: Planning, reasoning, execution, and memory
• Types of agents: Reflex, model-based reflex, goal-based agents
• Agent building: Python (LangChain with Azure) and no-code (Azure AI Agent Service)

Deploying a Generative AI Agent to automate customer support for an e-commerce platform. The AI will manage queries, process return requests, and provide personalized recommendations, covering deployment, optimization, containerization, and UI design.

Outcome: Learn the principles and practices of managing the lifecycle of LLMs, including model versioning, deployment strategies, and lifecycle management.

Topics Covered:

• Introduction to MLOps vs. LLMOps
• LLM lifecycle: Experimentation, training, deployment, monitoring
• Model versioning and registry in Azure Machine Learning
• Deploying LLMs as endpoints (Coding Track) or via PromptFlow (No-Code Track)

Sample Case Study: Automating healthcare documentation with LLMOps at HealthDocs

Outcome: Learn to monitor, log, and track LLM performance, implement CI/CD pipelines, and manage governance and security for scalable AI solutions.
Topics Covered:

• Monitoring performance: Latency, throughput, accuracy
• Logging and observability for LLMs in Azure
• CI/CD pipelines (Coding Track) and automated deployments with PromptFlow (No-Code Track)
• Security best practices for Azure LLM deployments

A digital marketing agency using LLMs to automate ad copy, blog posts, and social content needs to adopt LLMOps practices for monitoring, logging, and model management to ensure scalability and quality.

In this masterclass, you’ll learn to overcome standalone model limitations by incorporating external context, understand the MCP architecture (Host-Client-Server) and its components, and explore real-world benefits through hands-on exercises.

Outcome:

• Understand common terminologies in NLP (Corpus, Vocabulary, Out-of-Vocabulary)
• Explore text data and the need for NLP (purpose and challenges of working with text data)
• Learn text preprocessing methods (handling accented/special characters, lowercasing, stop word removal, tokenization, stemming & lemmatization)
• Apply text vectorization techniques (Bag of Words, Term Frequency–Inverse Document Frequency)
• Understand the ML NLP pipeline (text to prediction)
• Distinguish between base LLMs and chat LLMs
• Learn how large language models are trained
• Gain familiarity with the multi-step GPT assistant pipeline used in modern chat LLMs

Outcome:

• Introduction to machine learning concepts
• Understand supervised vs. unsupervised learning
• Differentiate between classification and regression
• Learn the role of evaluation metrics in ML
• Explore image-related terminology in computer vision (pixels, image channels, RGB images)
• Understand image processing techniques (edge detection, image filters, kernel-based feature extraction)
• Introduction to convolutional neural networks (CNNs): convolutional layers, multiple layers, padding, and pooling

Outcome:

• Introduction to Responsible AI: Definition and importance in today’s ecosystem; comparison with Ethical AI
• Core principles of Responsible AI: Fairness & inclusivity, transparency & explainability, accountability & governance, privacy & security
• Understand risks and vulnerabilities in LLMs: Limitations, bias and fairness issues, security vulnerabilities
• Explore mitigation strategies: Bias mitigation, hallucination control, reliability enhancement, safety/security measures, governance & compliance frameworks

Outcome:

• Introduction to prompt refinement and DsPy
• Core concepts of DsPy for refining and optimizing prompts