Senior Staff Inference Performance Engineer

What you'll do

• Diagnose and isolate performance issues across orchestration and inference-serving layers; fix issues or hand them off with actionable evidence.
• Protect throughput, latency, and stability by closing performance gaps against bare metal and capturing workload requirements in orchestration configuration.
• Build lightweight telemetry, benchmarks, and investigation playbooks; mentor engineers in profiling-first methods.

What they ask for

• 8+ years of hands-on performance or systems engineering, including resolving regressions across multiple stack layers.
• Deep systems fundamentals including Linux internals, NUMA, CPU pinning and cpusets, PCIe, memory and cache behavior, and network paths.
• Experience running performance-sensitive workloads on Kubernetes; working knowledge of LLM inference serving.
• Proficiency in Python and C/C++; fluency with AI and agentic tools, with discipline to verify their output against real data.

Nice to have

• Experience with LLM serving frameworks such as vLLM, SGLang, or TensorRT-LLM.
• Hands-on experience with speculative decoding, disaggregated or heterogeneous serving, GPU programming and profiling, or Rust.

Pay

$175K – $265K (base salary)

Tools

Kubernetes, Linux, Python, C/C++, vLLM, SGLang, TensorRT-LLM, Rust

AI focus

Performance engineering for LLM inference workloads, including serving, speculative decoding, and attention-FFN disaggregation; use AI and agentic tools for investigation, code search, log triage, and drafting instrumentation.