I've (@Isaac) been doing a variety of tasks in support of our topic infrastructure (with an eye towards our recommender systems and supporting campaigns) that will be captured here.
Topic infrastructure is a catch-all to capture the ecosystem of filters we might provide for subsetting content, especially within our recommender systems. This currently includes the ORES topics but will hopefully be expanded to include countries, Wikidata-based people tags, WikiProjects, quality scores (and potentially related features like image/reference counts), and perhaps others based on needs.
Some expected tasks:
- Formalize vision sketched out in https://docs.google.com/document/d/1qp-NPbP1pT7S2_VC9wCMIndpDEzMSPj1c62eOcTxbjc/edit?usp=sharing
- Support hypothesis generation during Annual Planning
- Build buy-in from the various teams that would use these features (Campaigns, Language, Growth, Apps, Community Growth)
- Coordinate with teams that this would depend on (ML Platform, Search)
- Guide technical changes to standardizing existing recommender systems to better use the topic infrastructure hosted in the Search index.