Implement Generative AI engineering with Azure Databricks (DP-3028-A)

Participants should have:

• Familiarity with core artificial intelligence and machine learning concepts
• Experience working with Azure Databricks environments
• Understanding of data engineering or data science workflows
• Exposure to Python or similar programming languages is recommended

Target audience

This course is designed for:

• Data scientists building advanced AI models
• Machine learning engineers operationalising AI systems
• AI engineers developing generative AI applications
• Technical professionals seeking to scale AI solutions using Azure Databricks

By the end of this course, learners will be able to:

• Explain generative AI engineering concepts within Azure Databricks
• Design and implement retrieval-augmented generation architectures
• Apply multi-stage and agent-based reasoning techniques in AI workflows
• Fine-tune large language models for domain-specific tasks
• Evaluate generative AI systems using modern performance and quality metrics
• Apply responsible AI principles to ensure ethical and compliant solutions
• Manage and operationalise generative AI solutions using LLMOps practices

Fundamentals of generative AI and large language models

• Overview of generative AI and its role in modern AI platforms
• Understanding large language models and transformer architectures
• Common use cases for enterprise generative AI solutions
• Key challenges in deploying generative AI systems at scale

Using Azure Databricks for generative AI workloads

• Introduction to Azure Databricks as a unified analytics platform
• Leveraging Apache Spark for distributed AI workloads
• Managing data pipelines for generative AI applications
• Integrating Databricks with Azure AI services

Retrieval-augmented generation architectures

• Principles of retrieval-augmented generation
• Combining vector search with language models
• Designing pipelines for contextual data retrieval
• Improving response accuracy and relevance with external knowledge sources

Multi-stage and agent-style reasoning patterns

• Understanding multi-step reasoning in generative AI systems
• Designing agent-based workflows using large language models
• Orchestrating tools and APIs within AI pipelines
• Enhancing decision-making through chained reasoning approaches

Fine-tuning large language models

• Overview of fine-tuning techniques and approaches
• Preparing datasets for supervised fine-tuning
• Parameter-efficient tuning methods
• Evaluating improvements from fine-tuned models

Evaluating generative AI systems

• Key evaluation metrics for large language models
• Automated and human-in-the-loop evaluation strategies
• Detecting bias, hallucinations, and model drift
• Benchmarking and continuous improvement practices

Responsible AI and governance considerations

• Principles of responsible AI in generative systems
• Managing risk, bias, and ethical concerns
• Ensuring compliance with organisational and regulatory standards
• Implementing governance frameworks for AI solutions

Managing generative AI solutions with LLMOps

• Introduction to large language model operations
• Versioning, monitoring, and lifecycle management of models
• Deploying generative AI applications in production
• Scaling and maintaining AI systems using Azure Databricks

Exams and assessments

There are no formal exams included in this course. Learners will complete knowledge checks and practical exercises throughout the day to reinforce key concepts and validate understanding of generative AI engineering techniques.

Hands-on learning

This course includes:

• Guided labs using Azure Databricks for generative AI workflows
• Practical exercises in retrieval-augmented generation and model fine-tuning
• Scenario-based activities focused on real-world AI engineering challenges
• Instructor-led discussions on applying LLMOps in production environments