GPU Performance Engineer (CUDA/Triton)

Location: Redwood City, California, United States; hybrid

Company

hirify.global builds unified general intelligence systems that generate, understand, and operate in the physical world, with a focus on multimodal AI and vision.

What you will do

• Profile and optimize GPU, CPU, and accelerator code to maximize utilization and minimize latency.
• Develop high-performance PyTorch, Triton, and CUDA kernels and custom operations.
• Build fused kernels using tensor cores and modern hardware features across platforms.
• Optimize transformer model architectures and implementations for distributed multi-node production deployment.
• Create performance monitoring, analysis, and automation tools.
• Research and implement optimization techniques for transformer models.

Requirements

• Expert-level Triton and CUDA programming with strong GPU optimization skills.
• Strong PyTorch experience, including kernel development and custom operations.
• Proficiency with profiling tools such as NVIDIA Nsight, torch profiler, and custom tooling.
• Deep understanding of transformer architectures and attention mechanisms.
• Ability to work in a hybrid setup in Redwood City, California.

Nice to have

• Experience with compilers and exporters such as torch.compile, TensorRT, ONNX, or XLA.
• Experience optimizing inference workloads for latency and throughput.
• Knowledge of Triton compiler and kernel fusion techniques.
• Knowledge of warp-level intrinsics and advanced CUDA optimization.

Culture & Benefits

• Work on multimodal models intended to support general intelligence.
• Equal opportunity employment environment.