Etched
Performance Modeling Engineer
$175K–$370K
San Jose · On-site · Full-time · Posted Apr 21, 2026
About the job
Develops and validates performance models for Etched’s accelerator architecture, analyzes deep learning workloads, and guides hardware/software co-optimization and future architecture decisions.
What you'll do
- Build performance models and projections across workloads and configurations; validate models against systems and silicon.
- Profile workloads to find micro-architectural bottlenecks and identify optimization opportunities.
- Evaluate system and silicon options to inform architectural decisions.
What they ask for
- Experience building analytical- or simulation-based performance models, or strong performance modeling and analysis skills.
- Solid understanding of computer architecture and micro-architecture, particularly for accelerators.
- Experience profiling and analyzing deep learning workloads on hardware accelerators.
- Solid software engineering fundamentals, with an eye toward auditability and maintainability.
Nice to have
- Deep knowledge of GPU architectures and/or programming models like CUDA.
- Experience mapping models to multi-chip inference systems; familiarity with transformer architectures, inference serving optimizations, simulators, or accelerator development is also noted.
Pay
USD $175K – $370K base salary
Tools
CUDA, gem5, GPUs, TPUs, ASICs, FPGAs, CGRA
AI focus
Performance modeling and optimization for inference-focused AI accelerators and deep learning workloads.
Model-written summary · the original posting is authoritative