perit.ai
EngineeringNewText

CUDA Engineering Expert

Review CUDA kernel implementations, audit GPU code for correctness and performance, and evaluate AI-generated systems code for safety-critical inference paths.

Rate
$130–$200 / hr
Placed this month
12
Commitment
6–12 hours per week, fully remote
Responsibilities

What you'll do

  • Review CUDA kernel code for correctness, memory safety, and performance.
  • Identify race conditions, memory leaks, and numerical instability in GPU code.
  • Evaluate AI-generated systems code against production engineering standards.
  • Benchmark kernel performance and recommend optimization strategies.
Requirements

What we need

  • 5+ years of GPU programming experience (CUDA, ROCm, or equivalent).
  • Production experience with ML inference optimization.
  • Strong C/C++ and systems programming background.
  • Experience with profiling tools (Nsight, nvprof, or equivalent).
Ready to apply

15 minutes. One interview. Verified for life.

Complete one domain interview and you'll be matched to paid research work — including this role — within a week.