Senior Modeling Architect, Performance Benchmarking

Neurophos • Austin, TX

22h • $210,000 - $250,000 • Onsite

About The Position

We are seeking a performance engineer to own the benchmarking numbers behind the T100 optical inference accelerator. Architecture and product decisions here are made on measured performance and energy, and this role produces those figures for the same workloads at every level of fidelity we use: roofline and limiter analysis, architecture performance models, in-house RTL simulation, and measured runs on competing GPUs and accelerators.
You will join the Architecture and Modeling team, set the measurement methodology, and keep it current as models, software stacks, drivers, and hardware generations turn over. Every result ships with the harness, config, plots, logs, and assumptions behind it, so anyone can rerun it and see how the number was reached.

Requirements

• BS or MS in Computer Engineering, Electrical Engineering, Computer Science, or equivalent practical experience.
• 5+ years of experience in GPU performance engineering, accelerator benchmarking, HPC performance measurement, or ML systems measurement.
• Track record of building or operating benchmark harnesses that produced measured results on real GPUs or accelerators, including turning a Hugging Face model card, paper, or application description into a runnable benchmark.
• Hands-on experience with roofline analysis, limiter analysis, or analytical performance modeling.
• GPU performance analysis with NVIDIA Nsight Systems and Nsight Compute, or an equivalent profiler, covering HBM-bound versus compute-bound analysis, precision (FP16, BF16, FP8, INT8), and batching.
• Working knowledge of LLM inference stacks such as Hugging Face, vLLM, SGLang, or TensorRT-LLM, including prefill versus decode, continuous batching, and MoE.
• Proficiency in Python for harnesses, parsing, and plots, and comfort working in Linux.
• Cloud GPU operations on AWS, GCP, or Azure, including containers, instance types, drivers, quotas, and cost.

Nice To Haves

• Experience correlating a performance model or RTL/Verilator simulation against measured silicon or GPUs.
• GPU kernel work in CUDA, CUTLASS, or Triton, or familiarity with PyTorch internals.
• Familiarity with current inference-serving internals such as PagedAttention, FlashAttention, speculative decoding, and disaggregated prefill.
• Distributed inference experience covering collectives, all-reduce, NCCL, NVLink, and InfiniBand, or work with MLPerf or production inference benchmarking pipelines.
• Background at a hyperscaler, GPU vendor, accelerator company, or inference lab.

Responsibilities

• Own the performance and energy metrics that architecture, product, and leadership rely on, and keep them consistent across modeling fidelities and measured hardware.
• Produce numbers for the same workloads across all fidelities we use internally: roofline and limiter models, architecture performance models, in-house RTL simulation, and measured competitor hardware. Work with the modeling team to keep the simulated and measured workload sets aligned.
• Hold workload definitions constant across fidelities, including model or application, sequence length, batch, precision, prefill versus decode, and tensor, pipeline, and sequence parallelism.
• Bring up inference workloads from Hugging Face, PyTorch, published papers, and vendor stacks such as vLLM, SGLang, TensorRT-LLM, and Triton Inference Server. These include dense and Mixture-of-Experts (MoE) transformers, attention and KV-cache reduction strategies, and hybrid/SSM models, as well as retrieval, speech, vision, and recommendation workloads that map onto the accelerator.
• Measure competing GPUs and accelerators end-to-end, owning the cloud or lab account, image, drivers, and run recipe.
• Report time to first token (TTFT), inter-token latency (ITL), tokens per second, tokens per second per watt, and energy, using nvidia-smi, DCGM, power capping, or equivalent instrumentation.
• Document where RTL simulation, the performance model, and measured competitor results disagree, and attach the configs and logs behind each.
• Maintain a reviewed internal benchmark suite. Keep internal-only results clearly separate from anything cleared for customer or public use, and route external claims through the designated approver before they ship.

Benefits

• 100% coverage of base health plan premiums for you and your dependents, plus HSA contributions.
• Unlimited PTO.
• 401(k) matching and stock option opportunities.
• Full suite of voluntary benefits, including Dental, Vision, Life, Hospital, Critical Illness, and Accident insurance.
• Personalized Benefits. Choose the plans that fit your life and take the cash back for those that don’t.

What This Job Offers

• Job Type: Full-time
• Career Level: Senior
• Education Level: Associate degree