Description

Design, implement, and maintain highly-optimized communication runtimes for Deep Learning frameworks and HPC interfaces on GPU clusters.

This role is on-site.

Responsibilities

• Develop system software enabling interactions among GPUs and between GPUs and other system components.
• Contribute to parallel programming interface specifications like MPI and OpenSHMEM.
• Create proof-of-concepts to evaluate extensions in programming models and new hardware features.
• Maintain communication runtimes for frameworks such as NCCL in TensorFlow and PyTorch.

Required Skills

• 5+ years of relevant experience in systems software or HPC.
• Excellent C/C++ programming and debugging skills.
• Strong experience with Linux environments.
• Expert understanding of computer system architecture and operating systems.
• Experience with parallel programming interfaces and communication runtimes.
• M.S./Ph.D. in CS/CE or equivalent experience.

Preferred Skills

• Experience with CUDA programming and NVIDIA GPUs.
• Knowledge of high-performance networks like InfiniBand or iWARP.
• Experience with Deep Learning Frameworks (PyTorch, TensorFlow) or HPC applications.

Key Skills

Education

Any Graduate