As part of product convergence within Match Group, several brands (including Meetic) are gradually sharing common platforms: technical infrastructures, user features, and business objectives. This unification is not only technical — it raises a concrete question: what to do with the AI ​​models developed separately by each app? Should we choose, merge, or coexist?

This talk offers feedback from the field on the dilemmas encountered when merging these ML stacks: how to compare models optimized for different contexts, arbitrate between performance and product alignment, manage inter-app cold-start, or maintain quality while unifying the pipeline

‍