Machine Learning & Production Readiness – Why Operations of ML is different

19.03. | 10:15 - 10:45 | tech.stage
Testing and monitoring are essential for stable machine learning solutions in operations. So far, however, only a few companies have transferred machine learning prototypes into operational use. Accordingly, there is little experience as to which tests are important and how exactly monitoring should take place. It shows some best practices of the ML pioneer Google. Learn how to design tests for ML applications in the areas of data, model, infrastructure and monitoring and what is meant by „hidden technical debt in machine learning systems“.
Jacqueline Bloemen (BARC GmbH)