Description

Own performance characterization, root-cause analysis, and tooling for large multi-GPU and multi-node clusters.

This role is hybrid.

Responsibilities

• Analyze performance interactions between libraries and hardware (GPU, CPU, Networking) and software components.
• Evaluate proof-of-concepts and conduct trade-off analysis for competing technical solutions.
• Triage and root-cause performance issues reported by customers across the entire stack.
• Collect performance data and build visualization tools and infrastructure for analysis.
• Implement micro-benchmarks in C/C++ and modify existing codebases as required.

Required Skills

• 5+ years of experience in performance engineering and HPC.
• Strong parallel programming skills with experience in MPI, NCCL, UCX, or NVSHMEM.
• Deep understanding of computer system architecture, hardware-software interactions, and OS principles.
• Proficiency in C/C++ and Python for scripting and benchmarking.
• Experience debugging performance issues on large-scale HPC clusters.
• Familiarity with containers and scheduling tools including Kubernetes, SLURM, Ansible, and Docker.
• M.S. or PhD in Computer Science or related field.

Preferred Skills

• Practical experience with Infiniband/Ethernet networks, RDMA, topologies, and congestion control.
• Experience debugging network issues in large-scale deployments.
• Familiarity with CUDA programming, GPUs, and Deep Learning Frameworks such as PyTorch or TensorFlow.

Key Skills

Education

Any Graduate