About Me
I'm a senior SWE in the applied machine learning (AML) team of Bytedance Inc. I work on building scalable and efficient machine learning systems and platforms to support DNN model training and serving. Our infrastructure supports most of Bytedance's neural network models including video recommendation, ads, search and e-commerce. Within AML, I led several projects around elastic resource provisioning for global serving and communication efficient training on GPUs.
Before that, I was a research scientist in the network infrastructure team working on software-defined networking projects. I led the projects on data center and WAN (wide-area network) bandwidth resource allocation and flow scheduling for Bytedance's global network infrastructure.I obtained my PhD degree from the Department of Computer Science, Rice University, advised by Prof. T. S. Eugene Ng. I also worked closely with Prof. Ang Chen.
During my PhD studies, I worked on problems that revolve around the performance, fault-tolerance, and reliability issues of modern datacenter networks. I'm interested in exploiting emerging underlying technologies such as optical devices and programmable switching ASICs to solve those problems. Contact Info: dingming.wu AT bytedance.com