Agentic AI Systems
CourseFreeThe Diploma in Agentic AI Systems is a comprehensive, self-paced program designed for beginners and professionals who want to master the theory, engineering, and deployment of modern autonomous and semi-autonomous AI agents.
Students will explore how AI agents perceive, reason, plan, and act in dynamic environments. The course covers foundational concepts, cognitive architectures, agent workflows, memory management, planning strategies, tool integration, multi-agent collaboration, and governance.
Learners will gain hands-on experience with popular agent development frameworks (LangChain, AutoGen, CrewAI, Semantic Kernel), learn professional safety and evaluation techniques, and develop practical skills to deploy agents in real-world business, research, and productivity contexts.
The diploma culminates in a capstone project, where students design, build, and evaluate a full agentic system, applying all concepts learned in a professional-grade, industry-aligned setting.
Graduates of this course will be prepared to build reliable, safe, and sophisticated autonomous AI agents, making them highly employable in AI engineering, automation, research, and enterprise roles.
- 1Understanding cognitive load & how LLM agents actually “think” differently from humans
- 2Recommended software stack (non-code + coding paths)
- 3How to maintain a research notebook (industry practice)
- 41.1 AI Foundations
- 51.2 LLM Cognitive Primitives
- 61.3 Agentic Paradigms
- 7Practical Lab: Analyze and compare three different agent behaviors in a constrained environment.
- 82.1 Memory Architecture
- 92.2 Tool-Use Mechanics
- 102.3 Reasoning & Planning
- 11Practical Lab: Build a tool-augmented agent and test failures, recovery, and decision logs.
- 123.1 Classical Agent Architectures
- 133.2 Modern Agent Orchestration
- 143.3 Multi-Agent Systems (MAS)
- 15Practical Lab: Design a multi-agent debate framework with scoring, arbitration, and logging.
- 164.1 Technical Depth: LangChain, CrewAI, AutoGen, Semantic Kernel
- 174.2 Orchestration engines (Temporal.io, AWS Step Functions)
- 184.3 Code-Lab Track: Build your first agent & multi-agent system
- 195.1 Agent Goal Specification
- 205.2 Prompt Engineering for Agents
- 215.3 Stability Design Patterns
- 225.4 Failure Handling
- 23Lab: Redesign a failing agent to achieve 30–50% higher task reliability.
- 246.1 Knowledge Agents
- 256.2 Autonomous Productivity Agents
- 266.3 Creative & Generative Agents
- 276.4 Enterprise Agents
- 28Lab: Build a domain-specific enterprise assistant with logging and safe action execution.
- 297.1 Technical safety (alignment, reward hacking)
- 307.2 Ethical frameworks applied to agents
- 317.3 Risk scoring & Governance policies (EU AI Act, NIST)
- 32Lab: Plan a safety strategy for a hypothetical enterprise deployment.
- 338.1 Evaluation Theory
- 348.2 Real-World Reliability
- 358.3 Agent Benchmark Suites
- 36Lab: Create a custom evaluation pipeline for an agent.
- 379.1 Proposal & System Design
- 389.2 Implementation Phase
- 399.3 Logs, Evaluation & Risk Analysis
- 409.4 Final Defense
- 4110.1 Future LLM architectures
- 4210.2 Hybrid agents & Cognitive architectures
- 4310.3 Agent-based economies
- 4410.4 Emerging Job roles in AI agent ecosystems