Senior Middleware Software Development Engineer
Core
Develop and optimize the oneCCL Collective Communications Library for Intel GPUs and CPUs to enable high-performance scientific computing and machine learning on supercomputers like Aurora.
Role type
Senior IC software development engineer (HPC communication libraries)
Builds
oneCCL library, performance optimizations for collective communications
Domain
High-Performance Computing (HPC), Distributed Systems, Supercomputing
Deliverable
production ML models | product features
Required skills
C/C++ programming, Linux environment development, multithreaded programming, parallel computing, distributed computing systems, HPC communications libraries, collective communications libraries (MPI/oneCCL/NCCL), GPU software development, network communications stack development
Preferred skills
Ph.D. in CS/Engineering, performance optimization experience, complex debugging across hardware/software stacks, Intel architecture optimization, supercomputing environment experience, scientific computing patterns
Technologies
oneCCL, MPI, NCCL, SHMEM, Intel GPUs, Intel CPUs, Linux
Responsibilities
Design and maintain advanced features for oneCCL, optimize software for low latency and high bandwidth, collaborate with scientists on Aurora supercomputer, partner with hardware teams for software-hardware integration, conduct performance analysis and benchmarking, debug complex hardware/software stack problems
Seniority
Senior, hands-on IC
