Session: Beyond the AI Hype: Building Durable Agentic AI Applications
AI is evolving faster than any engineering discipline in history. Each week brings new frameworks, agent orchestrators, LLM tooling stacks, and “revolutionary” platforms, all claiming to be essential. The result is familiar: technologists feel paralyzed, leaders struggle to separate signals from noise, and teams waste months experimenting without a clear path forward.
Drawing on experience from high-pressure AI hackathons and real-world system building, this session uncovers the principles that allow expert builders to cut through the noise, and design architectures that endure as the landscape changes. We’ll explore when simple LLM patterns are sufficient, when agentic workflows become transformative, and how to recognize the signals that a prototype is becoming fragile, before it’s too late.
Attendees will gain the mental models practitioners use to make durable decisions in a fast-moving environment. We’ll break down how to scope an AI feature, determine a minimum viable architecture, reason about orchestration layers, and identify the inflection point where multi-agent patterns become essential.
By the end, engineers will know how to start building without drowning in choices, leaders will understand how to guide teams through uncertainty, and educators and founders will leave with a clear, expert approach to navigating AI’s accelerating complexity. This talk equips the audience not with a stack, but with clarity, confidence, and an enduring framework for thinking in an era defined by constant AI innovation.
