Senior Software Engineer - Perf and Benchmarking
Core
Building and maintaining a planet-scale performance data warehouse for ingesting, storing, and analyzing performance events across global data centers to support industry-leading MLPerf benchmarking publications.
Role type
Senior IC software engineer (performance & benchmarking)
Builds
Kubernetes-native benchmarking services and end-to-end MLPerf Training/Inference workflows
Domain
Cloud infrastructure, high-performance computing, AI/ML benchmarking
Deliverable
production ML models | infrastructure
Required skills
Python, Go, C++, distributed systems, high-performance computing, Kubernetes, CI/CD, observability tools, networked systems, performance fundamentals
Preferred skills
time-series databases, LSM-based storage engines, MLPerf frameworks, CUDA kernels, NCCL/SHARP, RDMA/NUMA, GPU interconnect topologies, llm-d, vLLM, PyTorch
Responsibilities
Develop and enhance Kubernetes-native benchmarking services measuring latency, throughput, jitter, and cost-per-request; Implement and maintain benchmarking workflows for MLPerf runs including workload setup and result validation; Participate in architecture design discussions; Break down engineering tasks into milestones and deliver high-quality code; Collaborate to maintain reproducible, well-documented benchmarking processes; Provide code reviews and mentor junior engineers
Seniority
Senior, hands-on IC