Coursera Flash Sale
40% Off Coursera Plus for 3 Months!
Grab it
This 20-minute Open Compute Project talk by Gustavo Castellanos (Meta - Network Engineer) and Arushi Sharma (Meta - Hardware Analytics Engineer) explores an innovative framework for automating hardware failure diagnosis during New Product Introduction (NPI) phases. Learn how the team transformed the traditionally cumbersome process of manually analyzing millions of raw error logs into an efficient automated system. Discover how their solution integrates relevant tags to provide triage insights, identify trends, and detect patterns in errors. The presentation details how GenAI and machine learning techniques characterize uncategorized errors, delivering benefits including faster error categorization, improved test suggestions, reduced test flakiness, and standardized NPI readiness metrics. Gain valuable insights into methodologies that can significantly streamline hardware testing processes and reduce development delays.
