How We Built TLX: A Hardware-Native GPU Compiler for Production ML at Scale

3 points | by matt_d 10 hours ago