GPU Kernel Developer - Fully Remote

• Evaluate the quality and correctness of GPU/accelerator kernel development tasks for training and evaluating AI models.
• Assess numerical correctness, performance-benchmarking fairness, and task scoping across diverse kernel task types.
• Review compilation and runtime validity, providing clear, rubric-based written feedback.
• Collaborate with AI research teams to improve model training and evaluation processes.
• Work independently and asynchronously to meet deadlines and enhance AI model performance.

GPU (Graphics Processing Unit), Kernel Programming, Artificial Intelligence (AI), Benchmarking, Research Laboratory, Kernel Programming, GPU (Graphics Processing Unit), Performance Modeling, Time Management, CUDA (Compute Unified Device Architecture), Must-Have

• 3+ years of experience with CUDA, Triton, NKI, or Pallas (JAX).
• Strong understanding of numerical-correctness criteria for kernels.
• Experience with performance profiling and benchmarking (e.g., nsight, ncu).
• Familiarity with common compilation and runtime failure modes.
• Experience with at least three kernel task types.

Preferred

• Experience with both NVIDIA GPU and custom-accelerator ecosystems.
• Background in compiler engineering, MLIR, or intermediate-representation lowering.
• Understanding of memory-hierarchy optimization.
• Contributions to kernel libraries like cuBLAS, cuDNN, or Triton.

Mercor connects elite creative and technical talent with leading AI research labs. Headquartered in San Francisco, our investors include Benchmark, General Catalyst, Peter Thiel, Adam D’Angelo, Larry Summers, and Jack Dorsey.