DP-3028 Implement Generative AI engineering with Azure Databricks
Price Net € VAT €
Price Price on Request
Duration
1 day
Location
Course Language English
Training Solutions Online Live
Data platforms and generative AI are increasingly converging and shaping modern IT architectures. Azure Databricks provides a powerful foundation on which data-driven AI applications can be efficiently developed and operated.
Key topics
- Architecture of generative AI solutions in Azure.
- Scalable data processing with Azure Databricks.
- Integration and use of large language models.
- Prompt design, model adaptation, and evaluation.
- MLOps, security, and governance in the AI environment.
Prerequisites
Basic knowledge of cloud technologies, data processing, or software development, as well as a fundamental understanding of artificial intelligence, is expected.
Target audience
Aimed at professionals in data engineering, data science, AI development, and cloud architecture.
The focus is on stable, scalable, and responsible AI solutions that can be seamlessly integrated into existing cloud and data landscapes.
- Learn about the principles of generative AI
- Understanding how large language models (LLMs) work
- Recognize key components of LLM applications
- Use LLMs specifically for NLP tasks
- Understanding the key principles and structure of an RAG workflow
- Preparing data sets specifically for RAG processes
- Identifying relevant information using vector search
- Efficiently checking and optimizing results
- Fundamentals of multi-level reasoning systems
- Explore LangChain
- Explore LlamaIndex
- Explore Haystack
- Explore DSPy Framework
- Fundamentals of model fitting
- Preparation and structuring of training data
- Fitting and fine-tuning Azure OpenAI models
- Explore LLM evaluation in detail
- Analysis and evaluation of LLMs and AI systems
- Use of common metrics for LLM evaluation
- Explaining the use of LLMs as an evaluation tool
- Fundamentals of responsible AI
- Identifying risks
- Mitigating risks
- Using security tools
- Recognizing the shift from classic MLOps to LLMOps
- Understanding the basics of modern model deployment
- Using MLflow specifically for LLMOps
- Using Unity Catalog for secure model management
