Overview
This webinar explores the rapidly evolving field of agentic applications, where LLMs go beyond document summarization to initiate actions, trigger workflows, and achieve objectives autonomously. Dive into the core concepts of agentic AI applications, comparing them with traditional SaaS systems while examining powerful techniques like function calling and planning. Learn to understand tasks, workflows, and agency in agentic applications, determine when to build (or avoid building) agentic systems, and explore intent detection, function calling, and planning for enhanced task performance. Watch a comprehensive demonstration of an agentic app retrieving real-time information in RAG scenarios, performing actions like submitting forms, calling APIs, and modifying application states. The 56-minute presentation from Data Science Dojo covers everything from defining the agentic AI spectrum to infrastructure tools, function calling, and the role of Arch Gateway in agentic workflows, complete with code structure analysis and practical implementation examples.
Syllabus
0:00 Introduction to Agentic Applications
1:32 Defining "agentic AI" and its spectrum
5:48 Autonomous Agents
7:06 Configuration
15:48 Infrastructure & Tools
20:28 Function Calling
23:26 Role of Arch Gateway in agentic workflows
24:30 Demo of network device management agent
28:28 Code structure Python functions, FastAPI, YAML config
33:26 Live trace analysis
47:11 Q&A Highlights
55:51 Conclusion & Recommendations
Taught by
Data Science Dojo