Ryley Higa

Designed “Smart Monitors” an ML anomaly detection system that alerts customers based on irregularities in their log and metric data. Improved Sumo’s text-based clustering product for logs (logreduce) by modifying and distributing their algorithm over 100’s of nodes. My contribution increased the throughput (logs clustered per second) by 10x. Created an entity detection transformer for logs, which helped detect apps, services, and agents with few-shot examples. Increased frequency of code deploy by a factor of 3x by migrating our team’s services to Kubernetes. Created a new open source framework (Open Analytics) that allows customers to plug-in machine learning models into Sumo’s log search interface.