Ashish Vaswani

One of the most important architecture-level thinkers in modern AI, with influence spanning Transformers, efficient scaling, and mixture-of-experts systems.

A foundational transformer researcher whose work still matters because it connects the original architecture shift to later efforts on efficiency, scaling, and sequence modeling infrastructure.

A high-signal person to follow for the research arc from early transformer work into later sequence, vision, and multimodal model design.

One of the clearest people to study if you want the throughline from early neural sequence models to transformers, efficient long-context variants, and modern reasoning systems.

One of the clearest people to study if you want the throughline from early neural sequence models to transformers, efficient long-context variants, and modern reasoning systems.

A foundational transformer co-author who is now worth following for a very different reason: he is one of the few people trying to build a serious frontier lab around alternatives to the default scaling path.