Performance Modeling Engineer

About Etched

Etched is building hardware for frontier intelligence. We co-design chips, racks, software, and manufacturing to deliver best-in-class throughput and latency across both prefill and decode workloads. Our first products are heavily focused on inference

. Backed by hundreds of millions from top-tier investors and staffed by leading engineers, Etched is redefining the infrastructure layer for the fastest growing industry in history.

Key responsibilities

• Develop comprehensive performance models and projections for our architecture across varying workloads and configurations
• Profile and analyze deep learning workloads on our hardware to identify micro-architectural bottlenecks and influence optimization opportunities
• Drive hardware/software co-optimization by identifying where architectural features can unlock performance improvements
• Run regressions and validate performance models against real systems and silicon
• Inform next-generation architectural decisions by pathfinding across system and silicon options during design, proof-of-concept, and architecting phases
• You may be a good fit if you have at least one of the following: Strong performance modeling and analysis skills with experience building analytical-based or simulation-based performance models
• Solid understanding of computer architecture and micro-architecture, particularly for accelerators
• Experience profiling and analyzing deep learning workloads on hardware accelerators (GPUs, TPUs, ASICs, FPGAs, or others)
• Solid software engineering fundamentals with an eye toward auditability and maintainability
• Strong candidates may also have
• Deep knowledge of GPU architectures and/or programming models like CUDA
• Experience mapping models to multi-chip inference systems
• Familiarity with transformer model architectures and inference serving optimizations
• Experience with architecture simulators and performance modeling tools (gem5, trace-driven simulators, custom models)
• Exposure to ASIC, FPGA, or CGRA-based accelerator development and hardware/software co-design principles
• Published research in computer architecture, ML systems, or hardware acceleration