If you want to learn:
How do AI agents work and what makes them autonomous?
What is agentic workflow and how does it differ from traditional automation?
What are the five core techniques that power agentic AI systems?
How do LLMs make decisions and execute complex tasks in an agentic loop?
What is tool calling and how do AI agents use external tools?
What common pitfalls should you avoid when implementing agentic AI?
Then this lecture is for you!
This lecture provides foundational understanding of agentic AI and how AI agents work autonomously to execute complex tasks. You'll discover the five essential tricks behind agentic workflow systems: the illusion of memory, thinking and reasoning with LLMs, chaining large language models together, tool calling and tool use, and the agentic loop that enables agents to work iteratively toward goals. The session explains how prompt engineering and context engineering allow agents to make decisions, how AI systems interpret input and output to orchestrate workflows dynamically, and how tool invocation enables agents to interact with external tools and APIs. You'll learn why agentic workflows differ from traditional workflows, understand how autonomous AI agents maintain context without human intervention, and discover the "human trap" - a critical pitfall in agentic AI systems. This foundational lecture prepares you to implement agentic workflows, understand how multiple agents collaborate in multi-agent systems, and grasp how LLM-based agents automate complex workflows through intelligent decision-making and tool integration.