How to Design and Build Resilient Machine Learning Systems

Explore strategies for designing and building resilient machine learning systems in this 30-minute conference talk from MLOps World: Machine Learning in Production. Learn from Dan Shiebler, Head of Machine Learning at Abnormal Security, as he delves into the challenges of real-world ML implementation. Discover how to handle system failures, pipeline breakdowns, service outages, and unexpected user behavior. Understand the importance of graceful degradation and smooth failure handling in effective systems. Examine why ML models trained on clean data often struggle in production environments and how to address feature distribution shifts. Gain insights into detecting, mitigating, and overcoming production-specific risks, illustrated with examples from Abnormal Security's experience. Equip yourself with the knowledge to create ML systems that can withstand the complexities and uncertainties of real-world applications.

How to Design and Build Resilient Machine Learning Systems