Ashish Vaswani

Ashish Vaswani

First author of "Attention Is All You Need", the 2017 Google Brain paper that proposed dispensing with recurrence and convolution entirely and building a sequence model out of self-attention alone. The resulting Transformer architecture trained far better in parallel than its predecessors, which made scaling to billions of parameters practical, and it underlies essentially every large language model in use today. He later co-founded Adept and then Essential AI. (See also: Transformer, Large language model, Noam Shazeer)