Performance Engineer

Acceler8 Talent

Mountain View (CA)

On-site

USD 180,000 - 260,000

Full time

11 hours ago
Be an early applicant

A complete application in a minute

Tailored resume and cover letter, ready to send.

Get past ATS filters

Job summary

Acceler8 Talent partners with an early-stage AI company building software that learns from training and inference workloads to improve performance, power efficiency, and cost. You’ll join a small team owning core technology across kernels, runtimes, and model execution.

You’ll profile GPU workloads, optimize kernels, investigate fusion and memory movement, and build reproducible benchmarks that verify numerical correctness.

Qualifications

  • Experience optimizing GPU kernels and understanding bottlenecks
  • Strong fundamentals in GPU architecture and performance
  • Hands-on CUDA or comparable GPU/HPC experience
  • Ability to explain changes and measured results

Responsibilities

  • Profile GPU workloads to identify bottlenecks in compute, memory, synchronization, and communication.
  • Optimize existing kernels and develop new kernels where needed.
  • Investigate kernel fusion, memory movement, occupancy, and parallel execution.
  • Build reproducible benchmarks and verify numerical correctness.
  • Collaborate with serving infrastructure and applied ML teams to translate kernel improvements into workload performance gains.

Job description

I'm partnering with an early-stage AI company building software that learns from training and inference workloads to improve performance, power efficiency, and cost.

Founded by Google’s former TPU power architecture lead and Tesla’s former Dojo Supercomputer silicon design lead, the team is hiring an engineer to connect low-level GPU optimization with measurable improvements across the wider system.

You’ll join a small team with direct ownership of core technology, working across kernels, runtimes, and model execution.

What you’ll work on
  • Profile GPU workloads to identify bottlenecks in compute, memory, synchronization, and communication.
  • Optimize existing kernels and develop new kernels where needed.
  • Investigate kernel fusion, memory movement, occupancy, and parallel execution.
  • Build reproducible benchmarks and verify numerical correctness.
  • Collaborate across serving infrastructure and applied ML to translate kernel improvements into better workload performance and efficiency.
What we’re looking for

You’ve personally modified or developed GPU kernels and can explain the bottleneck you addressed, the changes you made, and the measured result.

You’ll bring strong engineering fundamentals, an understanding of GPU architecture, and hands-on CUDA or comparable GPU/HPC experience. Broader performance engineers with compiler, runtime, or systems depth are equally relevant, provided you’re comfortable changing kernels.

Experience with any of the following would be useful:

  • Triton, CUTLASS, cuDNN, PTX/SASS, or Nsight.
  • Roofline analysis, Tensor Cores, memory coalescing, or occupancy tuning.
  • NCCL, HBM, NVLink, NVSwitch, or multi-GPU communication.
  • ROCm/HIP, XLA/TPU, or optimization across different accelerator architectures.

You don’t need experience with every tool listed, or to have written kernels from scratch. Strong early-career engineers and exceptional graduates are welcome.

Get your free, confidential resume review.

or drag and drop your file here.

Similar jobs

Similar jobs worth comparing

GPU/Kernel Engineer
GPU/Kernel Engineer

Acceler8 Talent • Palo Alto (CA)

On-site
USD 180,000 - 240,000
GPU Performance Engineer | Experienced Hire
GPU Performance Engineer | Experienced Hire

SIG Susquehanna • Pennsylvania

On-site
USD 120,000 - 150,000
GPU Performance Engineer | Experienced Hire
GPU Performance Engineer | Experienced Hire

Susquehanna International Group, LLP • Bala Cynwyd (PA)

On-site
USD 100,000 - 130,000
Kernel Engineer
Kernel Engineer

Acceler8 Talent • San Francisco (CA)

On-site
USD 150,000 - 190,000
Software Engineer – GPU Kernel
Software Engineer – GPU Kernel

FriendliAI • San Francisco (CA)

On-site
USD 120,000 - 150,000
Flexible working hours
Daily lunch and dinner
Health check-up support
+3
GPU Kernel Performance Engineer — HPC/AI Infrastructure
GPU Kernel Performance Engineer — HPC/AI Infrastructure

Designworks Talent • Bellevue (WA)

Hybrid
USD 140,000 - 190,000
GPU Kernel Performance Engineer
GPU Kernel Performance Engineer

Acceler8 Talent • Mountain View (CA)

On-site
USD 180,000 - 260,000
GPU Kernel Engineer – CUDA, Triton & Accelerator Performance
GPU Kernel Engineer – CUDA, Triton & Accelerator Performance

Anyone AI • United States

Remote
USD 138,000 - 248,000
KERNEL ENGINEER
KERNEL ENGINEER

Makermaker • San Francisco (CA)

On-site
USD 180,000 - 240,000
KERNEL ENGINEER
KERNEL ENGINEER

Makermaker • San Francisco (CA)

On-site
USD 180,000 - 230,000
Open-source collaboration
Hybrid/On-site flexibility
Competitive compensation