You are leading a team of data analysts, scientists and engineers that are each members of cross-functional product teams where they work together with product managers, software and QA engineers as well as designers, if applicable. One of these product teams plans to discover an initiative to increase conversions of a particular user funnel by predicting a user's conversion action (dependent variable/label) based on various input/independent variables (features, e.g. from available session data).
Please describe a discovery process to create such an MVP and test the respective hypotheses. Include the various required steps including but not limited to model training, validation and testing, the people involved - including you as the data engineering manager - as well as any relevant technical resources and tools (e.g. software frameworks, model architectures, infrastructure etc.).
Imagine your company runs a large network of approx. 100K micro-websites with different kinds of consumer content. In total these websites see more than 100M visitors every month.
Please design a data model (e.g. using an entity-relationship diagram) for an ad exchange service that offers advertisers the ability to buy designated ad space to place their content on some of these pages to generate awareness, i.e. views/impressions, based on target audience definitions including geographic and content-related specifications. Besides the ad buying service, also consider tracking the campaigns' and their ad assets' performance.