Engineering and Production Techniques for Managing Feature Drift in AI Models

Explore engineering and production techniques for managing feature drift in large-scale AI models and systems in this 44-minute conference talk from the Toronto Machine Learning Series. Gain insights from Sharat Singh, CEO and Chief Architect at Quadrical.ai, as he addresses the challenges of non-cyclical feature drift in business environments. Learn about gradual and sudden changes caused by agent-environment interactions, and discover actionable best practices and system implementations derived from multiple case studies. Acquire valuable knowledge on how to effectively handle the evolving nature of data in AI applications and maintain model performance over time.