Staff AI Systems Performance Engineer

Requirements

• Experience with different Operating Systems (Windows, Linux, VMware).
• Scripting and/or programming languages, such as Shell scripts, Python, C/C++ are required.
• AI Infrastructure & Hardware Awareness - GPU/CPU architecture basics
• Experience in Performance & Benchmarking - profiling tools, system bottlenecks
• Debugging knowledge on performance bottlenecks in AI pipelines
• Experience in deriving data movement costs on Edge platforms
• Required:
• Bachelor's degree in Computer Science or Computer Engineering with 5+ years of experience, or a Master's degree in Computer Science or Computer Engineering with 2+ years of experience.

Benefits

Additional Information

The ideal candidate will be responsible for designing, defining, implementing, and enabling comprehensive benchmark tests for AI infrastructure platforms, including box-level GPU systems, multi-GPU servers, and GPU rack-scale deployments. This role requires a strong understanding of AI workloads, system architectures, and performance characterization methodologies across modern AI infrastructure environments. The individual will work closely with Marketing, Product Management and System Architecture teams to understand benchmark requirements and translate business and customer use cases into measurable performance validation strategies. The candidate will develop benchmark proposals, define evaluation methodologies, and execute performance studies for a wide range of AI applications, including Chat Assistants and LLM inference, Retrieval-Augmented Generation (RAG), Speech AI, Vision AI, multimodal AI, recommendation systems, and Image/Video Generation workloads. Responsibilities include selecting and optimizing benchmark frameworks, configuring AI software stacks, validating hardware and software performance, analyzing bottlenecks across compute, memory, storage, and networking subsystems, and generating detailed performance reports with actionable insights. The candidate will evaluate AI workloads across different hardware configurations such as GPUs, CPUs, accelerators, high-speed interconnects, NVLink/NVSwitch fabrics, storage architectures, and network fabrics to compare scalability, latency, throughput, power efficiency, and cost-performance metrics. The role also involves collaborating with internal and external partners to enable emerging AI models, benchmark suites, and infrastructure technologies, while ensuring reproducibility, automation, and continuous benchmarking capabilities within AI lab environments. Strong analytical, scripting, and performance tuning skills are essential, along with hands-on experience in AI frameworks, GPU computing, distributed inference environments, and performance monitoring tools. Essential Duties and Responsibilities: -Design & Validate AI Infrastructure for Benchmarks -Analysis of Benchmarks for End-to-end AI Infrastructure and develop test environment for Benchmark tests -Define and Perform Benchmark tests for AI workloads on storage systems -Evaluate GPU vs CPU vs storage bottlenecks in AI pipelines -Research and innovate Benchmarks for AI workloads on storage specific to Inference & training for major models -Design Benchmarks for Vector DB & KV cache -Optimize Data pipelines for Inference and training -Analyze Benchmark results, document and publish with recommendations

Your Match

How well this role fits your profile.

Company Intel

What employees say

Worked at Sandisk? Share your experience