Coursera Spring Sale
40% Off Coursera Plus Annual!
Grab it
Learn the harsh realities of deploying AI agents in production environments through this candid conference talk that goes beyond polished demos to reveal what actually happens when agents fail at 3 AM. Discover hard-earned lessons from deploying customer-facing agents handling sensitive financial data, including what breaks in production, how to prevent costly errors, and which patterns maintain stability in mission-critical systems. Explore architectural principles for reliability and observability, learn to design effective guardrails without compromising performance, and understand why development evaluation metrics often fail to predict real-world reliability. Gain practical insights into monitoring techniques that catch issues before customers notice them, architectural approaches that improve reliability at scale, and strategies for building functional guardrails that don't stifle agent capabilities. Understand the critical aspects of trust, risk management, and long-term system resilience when operating AI agents in production, all drawn from real-world experience and insights from the O'Reilly publication "Generative AI Design Patterns."
