Cloud Computing · Research interest
Cloud Infrastructure for Intelligent Applications
When local inference is worth the cost, and what breaks when retrieval systems get large.
Abstract
An interest in the operational tradeoffs of running intelligent applications on cloud and local infrastructure. This is not an active study.
Research questions
- When is local inference actually cheaper or safer than a cloud API?
- How should a team compare cost, latency, and failure in a hybrid setup?
Publication status
Research interest. No manuscript.