Senior Systems Software Engineer - Data Center Performance

Role overview

Assess how new computing platforms perform under demanding HPC and AI workloads, from individual nodes through racks and full clusters. You will combine hands-on experiments, profiling, and software engineering to find bottlenecks, verify fixes, and help teams deliver reliable performance at scale.

Responsibilities

- Benchmark nodes, racks, and clusters with microbenchmarks and representative applications; measure throughput, latency, and scaling. - Diagnose Linux, compute, memory, storage, and network issues, including scheduling, NUMA placement, resource contention, and communication overhead. - Correlate application profiles, system traces, hardware counters, and cluster telemetry to identify root causes. - Build focused reproductions and work with hardware, firmware, driver, library, and application teams to implement and validate improvements. - Develop Python tools for benchmark automation, data collection, and analysis; maintain repeatable baselines and detect regressions. - Validate new platforms and deployments, investigate underperforming nodes or links, and confirm that corrective actions restore workload performance.

Requirements

- Extensive hands-on Linux performance analysis, including processes and threads, CPU and memory architecture, NUMA, I/O, and networking. - Experience investigating multi-node HPC or GPU-accelerated systems and distinguishing local bottlenecks from communication, imbalance, and shared-resource contention. - Deep expertise in at least one area such as GPU application performance, Linux systems, high-speed networking, parallel storage, or HPC performance tools. - Proficiency in Python and experience developing or debugging performance-sensitive C, C++, and CUDA code. - Ability to design controlled experiments, interpret profiling and counter data, and demonstrate reproducible performance gains. - Clear technical communication and experience leading work across hardware and software teams; a relevant degree or equivalent experience and 8+ years of applicable experience.

Nice to have

- Experience with Nsight tools, Linux perf, eBPF, RDMA, MPI, or NCCL, and interconnects such as PCIe, NVLink, InfiniBand, or Ethernet. - Familiarity with Slurm, containers, cgroups, and the effects of workload placement and CPU, GPU, and NIC affinity. - Experience optimizing distributed training, inference, or scientific workloads, or building automated performance regression systems.