Track 4: Engineering Productivity Metrics
This track will investigate how we can customize typical software engineering metrics to usefully reflect progress for machine learning engineers. Not all their changes get deployed as they might not measurably push the final outcomes in A/B experiments. How can we still inform the engineers whether they go in the right direction?
Next to this, we will also focus on a shift-left of experimentation: Before deploying a change, an A/B test measures whether business metrics are improved by the change. We will develop tools to provide early feedback to the developers on the correct configuration of their experiments.
Track Leaders TU Delft
Lab Manager, Track Lead
Scientific Director, Track Lead
Track Leaders Meta
Track Lead
