Songlin Yang

Songlin Yang is a researcher at Thinking Machines Lab focused on pretraining and efficient language model architectures (linear attention, SSMs, GLA). She previously completed a PhD at MIT CSAIL and created the open-source Flash Linear Attention library.