Robust Anomaly detection at Netflix

Anomaly detection is the process of identifying data points that do not conform to normal behavior, and it is used ubiquitously at Netflix. To reduce false positives, identifying what is normal is critical. Once detected, analyzing the impact of the anomaly and prioritizing response is also critical for the system to succeed. In this session, we present a case study in payment processing at Netflix, where we use data science algorithms, scalable cloud infrastructure, and concise visualization to detect and take action on anomalies