Sounds like microservices deployed from a monorepo...
Which honestly may be the future if LLMs stay in a dev's toolkit. Plugging in an AI model to a monorepo provides so much context that can't be easily communicated across microservices in separate repos.