We all know that poor data quality leads to poor performance when it comes to training and serving AI models. But how much garbage is too much garbage?
In this session, researcher and ex-Epic Games data leader Andrew Koller will share how AI model performance continually degrades as unreliable data is introduced, and outline what measures teams can take to ensure that AI stays accurate, trustworthy, and valuable at each stage of the development lifecycle. You’ll also hear from Data Analyst and Monte Carlo expert, Elle Pyatt, on how to prioritize data quality incidents, learn best practices for creating and managing alerts, and identify metrics for improving incident response times to get all the benefits of alerts (without the alert fatigue).