From transformer to Terraform: Cohere’s large language model (LLM) takes flight
In the race to AI, it’s more crucial than ever to have strong, dependable infrastructure to build on. Since its founding in 2019, Cohere has experienced hypergrowth by empowering developers and enterprises to build amazing products while leveraging AI. Making these capabilities available to the world requires a tremendous amount of scale, both in Cohere’s LLMs and the underlying mass computational power. Come to this session featuring Sudip Roy, Director of Engineering at Cohere and Craig Alleva, Director of Customer Engineering, Google Cloud to discuss the passage to scale: from massive training on TPUs, to creating and operating a performant language model ecosystem served by Kubernetes and GPUs, and the technical and team choices made along the way.
Speakers: Sudip Roy, Craig Alleva
Watch more:
All sessions from Google Cloud Next → https://goo.gle/next23
#GoogleCloudNext