Join VultrVultr is seeking a highly skilled and experienced
GPU Performance and Benchmarking Engineer to drive performance validation and optimization of GPU infrastructure through rigorous benchmarking and systematic tuning of AI workloads. The ideal candidate is deep hands-on experience with GPU performance analysis, AI/ML workload characterization, and identifying and applying the parameters that maximize training and inference throughput. This is a highly visible role in a high-growth technology company, which will require strong analytical skills, familiarity with GPU profiling tools, and the ability to translate benchmark results into actionable tuning recommendations. This is your opportunity to join our fast growing team and leave your mark on Vultr and the future of Cloud Infrastructure.
Key Responsibilities- Design and execute performance benchmarks for AI training and inference workloads
- Profile and characterize GPU workloads to identify performance bottlenecks and optimization opportunities
- Systematically tune workload parameters (batch size, precision, parallelism, memory, etc.) to maximize throughput
- Establish and maintain performance baselines and success criteria across GPU platforms
- Develop benchmarking tools, scripts, and automation for repeatable performance validation
- Analyze and report benchmark results with actionable recommendations for engineering teams
- Validate performance of new GPU hardware platforms before production deployment
- Collaborate with GPU Engineers and Fabric Engineers to correlate performance with system-level health
- Track and evaluate emerging GPU architectures and software releases for performance impact
- Document benchmarking methodologies, tuning playbooks, and performance best practices
Qualifications- 3-7 years of experience in GPU performance engineering, benchmarking, or HPC
- Hands-on experience with GPU profiling and benchmarking tools (e.g., NVIDIA Nsight, DCGM, nccl-tests, MLPerf, GPU-Burn)
- Strong understanding of AI/ML workload performance characteristics (training vs. inference, batch sizing, precision modes)
- Experience tuning performance parameters across GPU, CUDA, and framework layers
- Familiarity with major GPU platforms (NVIDIA, AMD) and their performance tooling
- Proficiency in Python for benchmark scripting and data analysis
- Basic understanding of Linux systems and server hardware
- Familiarity with high-speed networking concepts (InfiniBand, RoCE, NCCL) and their impact on distributed performance
- Strong analytical skills with the ability to translate data into clear recommendations
- Excellent written communication skills for performance reports and documentation
Compensation$140,000 - $150,000
Final compensation will vary depending on years of experience, background/skill set, location, and applicable laws.
