Session: Architecting Intelligent Agents: Multi-Step Reasoning with Live Demo
AI is quickly evolving from chat-based interaction to systems that can think through tasks, make decisions, use tools, and take action. These intelligent agents represent a major step forward in how we design software, but building them effectively requires the right architecture, not just a powerful model.
In this talk, we’ll explore how intelligent agents work behind the scenes and what it takes to design them for multi-step reasoning. We’ll walk through the building blocks of agent systems, including planning, memory, retrieval, tool use, and execution, and discuss how these pieces come together to solve real problems. We’ll also cover practical lessons for creating agents that are reliable, scalable, and ready for real-world applications.
Through a live demo, attendees will see an intelligent agent process a realistic use case from start to finish. This session is ideal for developers and engineers who want to better understand modern AI agent architecture and learn how to apply these patterns in their own work.
Bio
Swarnamalya Mohan is a Senior Full Stack Developer at FedEx, focused on building scalable enterprise applications and intelligent technology solutions for logistics operations. She has experience across full-stack development, distributed systems, cloud technologies, and AI-driven applications, and is passionate about designing practical solutions that improve efficiency and user experience.