Agentic AI: System Design, Governance, and Innovation (AIB557)

This course provides an in-depth exploration of Agentic Artificial Intelligence, moving beyond foundational generative models to focus on the design, implementation, and oversight of autonomous AI systems. Students will delve into sophisticated agentic architectures, multi-agent collaboration patterns, and the frameworks required for robust system orchestration. A significant emphasis is placed on the practical challenges of building and deploying real-world agents, including tool use, planning, and continuous learning. The curriculum also provides a rigorous examination of agent assessment, featuring advanced evaluation benchmarks and performance metrics. Critically, the course is grounded in the principles of responsible AI, with a dedicated focus on governance, security, and ethical considerations, drawing upon global frameworks such as Singapore’s Model AI Governance Framework for Agentic AI. Upon completion, students will be equipped to design, manage, and critically evaluate complex AI agentic systems, enabling them to lead innovation responsibly in a rapidly advancing technological landscape.

• Introduction to Agentic AI: From LLMs to Autonomous Agents
• Agentic Architectures: Core Components and Design Patterns
• Frameworks for Agent Implementation (e.g., LangGraph, AutoGen)
• Multi-Agent Systems: Collaboration and Orchestration
• Advanced Planning and Reasoning for AI Agents
• Agentic AI Security: Safeguarding Autonomous Systems
• Assessment and Evaluation of Agentic Systems
• Human-in-the-Loop Design and Agent-Human Interaction
• The Business and Economics of Agentic AI
• Agentic AI Governance: Global Frameworks and Best Practices
• Singapore’s Model AI Governance Framework for Agentic AI
• Future of Work in an Agentic World

• Evaluate core principles of agentic AI architectures and design patterns
• Design solutions to address ethical, safety, and security challenges in autonomous AI systems
• Formulate assessment frameworks and performance metrics for agentic AI
• Construct multi-agent AI systems for complex problem-solving
• Design AI agents using modern frameworks and orchestration techniques
• Formulate and evaluate governance protocols for responsible deployment and management of agentic AI