Zhihao Jia

Compact student models can now achieve near EEG foundation model performance with significantly reduced model size and inference cost, thanks to a novel knowledge distillation framework.

Tensor program optimization just got a whole lot faster: Prism achieves up to 2.2x speedup over existing superoptimizers while *also* reducing end-to-end optimization time.

Unlock 2x faster LLM serving and slash warmup times by fusing kernels that gracefully handle dynamic shapes and data dependencies.