GenAIOps Enablement with Red Hat AI Enterprise (AI501)

Cílová skupina

This experience demonstrates how individuals across different roles must learn to share, collaborate, and work toward a common goal to achieve positive outcomes and drive generative AI innovation.

It is especially valuable for:

• AI Platform Users: AI engineers, application developers, data scientists, and data engineers building generative AI applications
• AI Platform Providers: ML/GenAIOps engineers and platform engineers deploying and managing AI infrastructure
• AII Platform Stakeholders: Architects and IT managers evaluating and overseeing generative AI adoption strategies

The scenario incorporates technical aspects of working with large language models and generative AI systems, offering practical insights into how these roles can align their efforts.

Struktura kurzu

Core Foundations

• GenAI Fundamentals

  Explore what GenAIOps is and how large language models work, including tokenization, context windows, and the factors that affect model behavior and performance.

• Experimenting with Prompts

  Learn to craft effective prompts using system prompts and user prompts, configure temperature and output parameters, and optimize prompts for specific use cases.

• Evaluating Your First AI-Enabled Application

  Implement prompt versioning, build evaluation pipelines, automate testing, and measure application quality systematically.

• Introducing the Orchestration Layer

  Introduce an orchestration layer for building GenAI applications, deploy backend services, and implement GitOps practices for continuous deployment.

Advanced Topics

• Integration and Orchestration

  Deploy vector databases, build RAG pipelines for knowledge-enhanced applications, implement tool calling, and create autonomous AI agents.

• Safety and Observability

  Deploy AI safety guardrails, implement GenAI security practices, enable the three pillars of observability, metrics, logs, and traces.

• Modeling Techniques

  Explore small language models for efficient deployment and multi-modal model capabilities for handling diverse input types.

• Optimization and Deployment

  Apply quantization and compression techniques for improved performance, explore fine-tuning approaches, implement Models as a Service (MaaS), and bring it all together in a production deployment.