Course
digicode: AGAIDD
Agentic AI – Deep Dive
Course facts
- In-depth understanding of agentic AI, including differentiation from chatbots and RAG
- Architectures and functioning of AI agents (single-agent and multi-agent architectures)
- Systematic evaluation and optimization of agent-based systems
- Security and governance for agentic AI
- Understanding potential and challenges
Develop AI agents that plan tasks, make decisions, and execute actions independently. The course combines theory with practice for the direct implementation of agentic AI solutions using modern open-source and cloud tools. The course content specifically covers:
1 Introduction to agentic AI
Understand agentic AI in an enterprise context. Recognize how semi-autonomous AI systems plan, decide, and act—and how they clearly differ from classic chatbots and RAG approaches.
2 Fundamentals of agent-based systems
Develop the central concepts of AI agents: roles of LLMs, tools, memory, and planning. Analyze possible applications of agent-based solutions in business processes.
3 Business case & application scenarios
Assess when Agentic AI creates real added value. Identify suitable tasks, evaluate alternatives, and derive sound decision-making criteria for the use of agents.
4 Decision-making & planning
Apply modern agent patterns: ReAct, tool calling, multi-step reasoning, and memory and state management. Build robust decision-making logic for autonomous AI agents.
5 Architecture & Orchestration
Understand typical agentic AI architectures. Orchestrate tools, APIs, and workflows, taking into account scaling, costs, and latency in productive operation.
6 Implementation of AI agents
Develop functional AI agents with Python, open source, and public cloud tools (Azure) in the Agentic AI course. Integrate external tools, data sources, and enterprise systems hands-on.
7 Evaluation & Optimization
Systematically monitor and improve agent-based systems. Use monitoring, testing, and optimization strategies to increase goal achievement and reliability.
8 Security & Governance
Implement security and governance concepts for Agentic AI. Use access controls, prompt and action logging, red teaming, content filtering, and guardrails.
9 Advanced Concepts
Deepen your knowledge with multi-agent systems, agentic RAG, human-in-the-loop approaches, workflow and multimodal agents – as an integral part of the Agentic AI Deep Dive.
- Hands-on implementation
- Interactive theory units
- Peer-to-peer exchange
- Software engineers
- Data scientists, ML engineers, AI engineers
- IT architects (e.g., software architects, enterprise architects)
- Project and product managers with programming experience
- Basic knowledge of software development and Python programming required
- Experience with AI technologies and cloud platforms is an advantage, but not essential