Mayank DagaAdvanced Micro Devices
Mayank Daga leads the machine intelligence software group at AMD, where we develop highly optimized GPU software for deep learning. We also interact with groups developing deep learning frameworks and help them succeed in optimizing their frameworks for AMD processors. We are currently developing MIOpen – a fully open source, high performance library for GPU kernels tuned for Deep Learning.
Mayank spent several years working on performance optimization of scientific and irregular applications on AMD accelerators including developing an industry leading sparse matrix-vector multiplication algorithm when announced. Prior to AMD, he graduated with a Masters in Computer Science from Virginia Tech, where he was adjudicated the Outstanding Masters Student by the Dept. of Computer Science for his work on architecture-aware optimization strategies on various heterogeneous computing systems.