Inference Systems Performance Architect

Highlights

• Architect end-to-end inference performance from tokenization through decode and f
• Build replayable benchmarks capturing real agentic production traffic for serving.
• Define heterogeneous disaggregated inference with GPU prefill and RDU decode.

What they offer

• Equity
• 95% employee medical premium coverage, 77% for dependents
• Health Savings Account with employer contribution
• Dental, Vision and FSA options
• Disability, Life and AD&D insurance
• Headspace subscription and Gympass+ membership
• One Medical membership and Employee Assistance Program

Ideal candidate

A veteran performance architect who thrives on ambiguous, first-of-kind systems problems and sets technical direction. You care about honest, production-faithful measurement and turning data into hardware and capacity decisions.

Requirements

• B.S. in Computer Science, Computer Engineering, or related field
• 12+ years of experience in performance engineering
• Direct experience with LLM inference serving, continuous batching and KV caching
• Familiarity with inference simulation frameworks or agentic benchmarking
• Public technical voice through talks, writing or community presence