Performance & Kernel Engineering Jobs
Wringing performance out of hardware — kernels, compilers, Triton and low-level optimization for ML workloads. 63 open now, refreshed daily.
- open roles
- 63
- companies
- 19
- list salary
- 33 · $120K–$485K
- visa mention
- 16
- remote
- 2
Observed across current open postings, refreshed daily — not a survey. Salary band is drawn only from roles that publish a range. Salary breakdown →
Performance and kernel-engineering roles are the optimization core of this niche — compilers, custom kernels (often Triton), graph-level transforms, and the low-level work of making a given workload run measurably faster on given hardware. It is consistently the largest specialty on this board, which tracks the moment: every accelerator company and serving stack is now competing on efficiency. These roles favor people who profile before they optimize, who are comfortable at the IR/compiler layer, and who treat a benchmark regression as a bug to be bisected.
Hiring most for this specialty: Graphcore 13 · Tenstorrent 11 · Anthropic 7 · CoreWeave 4 · d-Matrix 4 · MatX 4 · see all who's hiring →