Senior Architecture & Performance Modeling Engineer, AI Inference Silicon
New
P
Positron AIAI inference silicon
United StatesFull-TimeSenior
Salary220,000 - 270,000 USD per year
Apply NowOpens the employer's application page
Job Details
- Experience
- 5+ years of experience in computer architecture, microarchitecture, or performance modeling for accelerators, CPUs, or GPUs.
- Required Skills
- PythonC++Rust
Requirements
- Have 5+ years of experience in computer architecture, microarchitecture, or performance modeling for accelerators, CPUs, or GPUs.
- Demonstrate experience using quantitative models to influence architectural decisions, compare design alternatives, or recommend changes.
- Understand pipelines, queues, parallel execution, memory hierarchies, data movement, and interconnects.
- Have hands-on experience building analytical, transaction-level, or cycle-level performance models.
- Be comfortable working with incomplete specifications, making assumptions explicit, and balancing model fidelity with iteration speed.
- Have strong Rust skills, or expert-level C/C++ skills and willingness to work in Rust day to day.
- Be proficient in Python for workload analysis, experiments, and visualization.
- Be able to read SystemVerilog RTL and verification results, reconcile them with architectural specifications, and calibrate models.
- Preferred: Experience with LLM inference, including attention, KV caches, quantization, mixture-of-experts workloads, and multi-chip parallelism.
- Preferred: Experience connecting workload behavior and software execution choices to hardware performance.
Responsibilities
- Partner with architects to turn design questions into modeling studies with clear assumptions, alternatives, metrics, and recommendations.
- Analyze compute engines, memory capacity and bandwidth, dataflow, DMA, command processing, on-chip interconnect, and chip-to-chip communication.
- Compare architectural options, explain performance sensitivities, and recommend changes within system, power, area, and implementation constraints.
- Evaluate representative AI inference workloads across prefill and decode, batch sizes, context lengths, and dense and mixture-of-experts models.
- Communicate findings, investigate unexpected results, and update studies as workloads and architectural choices evolve.
- Develop analytical, transaction-level, and cycle-level models using an appropriate level of detail.
- Maintain reusable models in the native Rust simulator and correlate predictions against RTL simulation, emulation, or silicon measurements.
- Maintain regression coverage and document model assumptions, limitations, and discrepancies.
View Full Description & ApplyYou'll be redirected to the employer's site