attention head AI jobs Alerts Saved
AI jobs › GPU & kernels AI jobs

GPU & kernels AI jobs

This area sits between models and hardware. A training run or inference service is often limited less by the maths than by how well it uses the chips: whether data moves between memory and compute fast enough, and whether many GPUs spend their time computing or waiting on each other.

Kernel engineers write and tune the low-level code for operations such as attention and matrix multiplication, fusing steps so data stays close to the compute. Compiler engineers work on the layers that turn a model's code into kernels. Distributed training engineers split models and data across many devices and tune the communication between them, often through NVIDIA's NCCL library, and performance engineers profile whole systems to find where the time goes. These roles sit at labs, chip makers and companies that train or serve models at scale, and expect C++, CUDA or Triton, and a good grasp of how GPUs work: memory hierarchy, bandwidth and parallelism.

We count roles whose postings mention CUDA, Triton, GPU kernels, kernel engineering, development or optimisation, NCCL, distributed training, or GPU performance or programming. Kernel development can also mean work on an operating system's kernel, so a few systems roles are counted here too.

656 open AI roles at 280 employers, from their own job boards. Showing the newest 100.

Follow GPU & kernels33 new this week

New GPU & kernels roles by email, the morning after we find them, in one email with your other alerts. Narrow it, or keep every role.

How often
We email a link to confirm it.
NVIDIA · 112Amazon · 36Cerebras · 12Together AI · 11Cohere · 10Sciforium · 8Nebius · 7Inferact · 6Waabi · 6Wayve · 6Centific Global Solutions, Inc. · 5Crusoe · 5Datadog · 5DigitalOcean · 5Nuro · 5

Get new GPU & kernels roles every Monday

A job alert: new GPU & kernels roles from employers' own job boards, every Monday. Unsubscribe in one click.

Related