Agentic AI Design Patterns for GenAI and Predictive AI
50mAdvanced2025-10-16
Authors
Thomas Erl
Course details
Predictive AI and generative AI provide clear business value, but each has its own limitations and risks. By leveraging the autonomous reasoning capabilities of agentic AI, we can now overcome many of those limitations and we can now mitigate many of the risks. In this course, LinkedIn Top Voice and best-selling author Thomas Erl covers a set of AI design patterns dedicated to how agentic AI can be effectively positioned and utilized as part of predictive AI and generative AI solutions. This can result in AI systems that are more autonomous, self-reliant, efficient, and accurate, as well as AI systems that are enhanced with regards to how both humans and applications can interact and integrate with them. This course, intended for AI/ML engineers and AI solution architects, aims to fill in the gaps between agentic AI and predictive/generative AI by showing how these technologies can be effectively used together.
Skills covered
Software DesignSoftware ArchitectureAI for Personal ProductivityAI Agents and Agentic SystemsAI Productivity and Everyday UseBuilding with AIBusiness Software and ToolsSoftware DevelopmentOne-Off
Concepts
Introduction
- Introduction
- What you need to know
- AI agents and AI systems
Knowledge and Reasoning Design Patterns
- Knowledge and reasoning design patterns
- Autonomous self-RAG
- Automated feature discovery and engineering
- Causal inference
Orchestration and Creation Design Patterns
- Orchestration and creation design patterns
- Agent-led parallelization
- Hierarchical generation
- Voting ensemble
Validation and Adaptive Learning Design Patterns
- Validation and adaptive learning design patterns
- Escalating evaluation
- Autonomous self-correction
- Proactive retraining
- Shadow model deployment
Conclusion
- Next steps