Architecture & Performance Modeling Engineer, AI Inference Silicon (Senior / Principal)
Core
Define architecture and performance of next-generation AI inference accelerators for large language models using analytical models and simulation before RTL implementation.
Role type
Senior/Principal Architecture & Performance Modeling Engineer (AI Inference Silicon)
Builds
Next-generation AI inference chips (accelerators) for LLMs
Domain
Semiconductor / AI Hardware / Computer Architecture
Deliverable
production ML models
Required skills
Computer architecture fundamentals, Performance modeling (analytical/transaction/cycle-level), Rust, Python, SystemVerilog RTL, LLM inference workloads (prefill/decode/MoE), Data movement & interconnects
Preferred skills
LLM inference experience (attention, KV caches, quantization), Multi-chip parallelism, Workload behavior analysis
Technologies
Rust, Python, SystemVerilog, Simulation frameworks
Responsibilities
Collaborate with chip architects to translate questions into structured modeling studies; Conduct architectural trade-off analyses across compute, memory, and interconnects; Evaluate hardware configurations to optimize performance within constraints; Analyze AI inference workloads to identify bottlenecks; Develop and maintain reusable performance models in Rust; Validate model predictions against RTL simulations and silicon measurements; Document modeling assumptions and results; Communicate findings to engineering teams.
Seniority
Senior/Principal, hands-on IC