Scope. Roles that name CUDA, ROCm, Triton, or GPU kernel authoring: occupancy, memory hierarchy, and accelerator throughput.
Not here. A role appears here only when NearMetal’s own published summary, responsibilities, or requirements name the technology — never on a guess.
Languages named on these postings. C++ (8), Rust (3), Assembly (1).
Hiring most into this area. Baseten (2), CoreWeave (2), Tenstorrent (2), Fireworks AI (1), OpenAI (1).
Software Engineer - GPU Kernels
view source ↗
Software Engineer - Model Performance
view source ↗
Senior Software Engineer - GPU Kernel Authoring & Optimization
view source ↗
Senior Software Engineer, Inference
view source ↗
Member of Technical Staff, Performance Optimization
view source ↗
Training Performance Engineer
view source ↗
Member of Technical Staff (AI Inference Engineer)
view source ↗
Senior Software Engineer , AI and Search Team - C/C++ or Rust
view source ↗
Acceleration Kernel Developer Lead
view source ↗
Sr. Engineer, Kernel Development and Optimization
view source ↗