SambaNova
Inference Systems Performance Architect
$245K–$325K
San Jose, California, United States · Full-time · Posted Aug 16, 2026
- Experience
- 12+ years in performance engineering.
- Education
- B.S. in Computer Science, Computer Engineering, or Related Field; M.S. or PHD preferred.
About the job
Architects end-to-end inference performance, developing workload capture, benchmarking, modeling, simulation, and profiling to guide serving optimization, capacity planning, and future systems and hardware.
What you'll do
- Set technical strategy and build capabilities for workload capture, agentic benchmarking, performance modeling, and simulation.
- Develop actionable tools to profile distributed inference pipelines and locate bottlenecks across host, accelerator, and fabric.
- Lead cross-functional technical efforts, advise customers and partners, and mentor senior engineers.
What they ask for
- B.S. in Computer Science, Computer Engineering, or Related Field.
- 12+ years of experience in performance engineering, with demonstrated technical leadership on large-scale, complex systems.
- Expertise in distributed-systems performance analysis, realistic workload generation and simulation, and performance modeling calibrated against real workloads.
- Ability to lead cross-functional work, mentor senior engineers, influence organizational direction, and deliver high-complexity work independently.
Nice to have
- M.S. or PHD in Computer Science, Computer Engineering, or Related Field.
- Direct LLM inference-serving experience, familiarity with inference simulation or agentic benchmarking, or a public technical voice on systems performance.
Pay
$245,000 — $325,000 USD (base salary)
Tools
LLM inference serving, continuous batching, prompt/KV caching, prefill/decode disaggregation
AI focus
Improves and models LLM inference serving, including agentic workloads and heterogeneous inference with GPU on prefill and RDU on decode.
Model-written summary · the original posting is authoritative