CoreWeave
Principal Engineer - Perf and Benchmarking
$206K–$333K
Sunnyvale, CA / Bellevue, WA +2 · Principal · Full-time · Posted Dec 18, 2025
- Location
- Sunnyvale, CA / Bellevue, WA
Bellevue, WA
Sunnyvale, California, United States - Experience
- 10+ years building distributed systems or HPC/cloud services.
- Visa sponsorship
- Applicants must meet one of the stated export-control eligibility conditions: be a U.S. person, be eligible to access the information without export authorization, or be eligible and reasonably likely to obtain required authorization. CoreWeave may decline to pursue export licensing.
About the job
Leads CoreWeave’s benchmarking and performance team, building its global performance data platform and delivering audited ML performance benchmarks and optimization results.
What you'll do
- Set the benchmarking strategy, lead and mentor the team, and establish governance for reproducible performance claims.
- Lead MLPerf Training and Inference submissions and coordinate GPU and model-stack optimization with NVIDIA and open-source communities.
- Build scalable Kubernetes-based benchmark automation and performance data systems, and report results across workloads, software releases, and hardware generations.
What they ask for
- 10+ years building distributed systems or HPC/cloud services, with deep expertise in large-scale ML training or similar high-performance workloads.
- Track record architecting or building planet-scale data systems.
- Deep understanding of GPU performance, model-serving stacks, and distributed training frameworks.
- Proficient with Kubernetes and ML control planes; excellent communication skills.
Nice to have
- Experience with time-series databases, LSM trees, or custom storage engines.
- Experience running MLPerf or equivalent audited benchmarks at scale; OSS contributions, large-fleet benchmarking, or relevant publications and talks are also valued.
Pay
$206,000 to $333,000 USD base salary
Tools
MLPerf, Kubernetes, SUNK, Kueue, Kubeflow, CUDA, NCCL, Prometheus
AI focus
Benchmarks and optimizes AI/ML training and inference workloads, including LLMs, diffusion, vision, and speech models.
Model-written summary · the original posting is authoritative