Or Ordentlich

Develops optimal quantization methods for large language model weights and matrix multiplication, advancing efficient inference for large-scale AI models.