Etched
Head of Inference Performance Visibility
$200K–$400K
San Jose · On-site · Director · Full-time · Posted Feb 19, 2026
About the job
Define and build performance visibility systems that correlate hardware, runtime, and workload telemetry across AI accelerators, hosts, and distributed inference clusters.
What you'll do
- Design system-level telemetry collection, counter models, and cross-layer event correlation.
- Build time synchronization, trace-alignment, and distributed performance analysis tools for multi-accelerator systems and clusters.
- Shape performance analysis engines and developer-facing tools, and lead cross-functional architecture spanning hardware and software.
What they ask for
- Deep experience building complex systems at the intersection of hardware and software.
- Personally envisioned and built substantial profiling, tracing, or observability systems.
- Deep systems programming expertise in C++ or Rust, with shipped low-level infrastructure close to hardware or runtime systems.
- Experience correlating distributed events and designing performance analysis or observability infrastructure.
Nice to have
- Experience with high-performance computing systems and large AI training clusters.
- Experience with hardware counter design, instrumentation strategy, and performance modeling for large-scale ML workloads.
Pay
$200K – $400K USD base salary
Tools
C++, Rust
AI focus
Build performance visibility and analysis infrastructure for AI accelerator systems and distributed inference workloads.
Model-written summary · the original posting is authoritative