Model Monitoring

With the recent introduction of MLOps to the industry, data scientists and ML engineers are now ready to build their foundation for machine learning life-cycle management with DevOps and CI/CD practices incorporated. But this is just the beginning of the long road ahead. Over the last 20 years, Google has learned that ML requires us to meet unique challenges – including monitoring, logging, and alerting for deployed ML models. And that’s the missing link between current MLOps and the real-world production ML system. In this talk, we explain how we’re making it, not just possible, but easy to monitor and manage the quality of deployed ML models. View all of the sessions from the Applied ML Summit, including innovation sessions with leading data scientists, technical tutorials and panel discussions on MLOps → https://goo.gle/3ckHNsw #GoogleCloudSummit ML301