Summary
Masoud Moshref is a Principal System & Software Architect with 11 years of experience designing high-performance networking for AI/ML at hyperscalers, currently driving system and software architecture at NVIDIA. He previously led congestion control and kernel-bypass networking efforts at Google for RDMA and IPU-based products, contributing to projects like PaLM training and multiple SIGCOMM/NSDI-era research prototypes. His PhD work at USC produced several influential systems for scalable network monitoring and resource allocation (vCRIB, DREAM, SCREAM, Trumpet), and he holds both Google PhD and USC Provost fellowships. At Barefoot Networks he pushed the boundaries of in-network telemetry and P4 programming—building INT, DDoS detectors, and a 3.2Tbps parameter server prototype—demonstrating an unusual mix of hardware-aware software design and protocol innovation. Based in Detroit, he blends deep research pedigree with hands-on production deployments, optimizing low-latency, high-throughput fabrics for real-world ML and storage workloads.
11 years of coding experience
13 years of employment as a software developer
Doctor of Philosophy (Ph.D.) Computer Engineering, Doctor of Philosophy (Ph.D.) Computer Engineering at University of Southern California
Ejeiee high school
MSc Information Technology Engineering, MSc Information Technology Engineering at Sharif University of Technology