Efficiency team
The Efficiency team works on inference speed, cost and infrastructure so that capable AI is available to everyone, not only those who can pay the most.
Fast inference
Lower latency so answers arrive while you are still thinking about the question.
Cost per task
Doing the same work with fewer resources, and passing the savings to users.
Reliable infrastructure
Systems that stay up as usage grows around the world.