4+ years of experience writing performant GPU kernels (CUDA, ROCm, Triton, or equivalent)
Strong understanding of hardware-level concepts like memory hierarchy and warp scheduling
Proficient in profiling tools (Nsight, ncu) with a disciplined approach to measurement
Proven track record of delivering kernel optimizations that improve system performance
Solid systems expertise, understanding kernel integration within larger frameworks
Ability to read and work with Python and C++ in relation to kernel development
Responsibilities
Write and optimize GPU kernels for various workloads including attention variants and custom activations
Profile workloads using hardware counters to inform kernel optimizations
Collaborate with research teams to co-design kernels and algorithms
Integrate optimized kernels into training and serving stacks, benchmarking performance
Maintain kernel quality as hardware and frameworks evolve
Share kernel-level expertise across the team to foster a collaborative environment
Benefits
Opportunity to work on cutting-edge GPU technology
Collaborative environment with research teams
Focus on impactful performance improvements
Exposure to a variety of accelerator architectures
Potential for open-source contributions in the field
Full Job Description
ABOUT THE ROLE
You'll write and optimize the GPU kernels and supporting systems software that makes our training and inference workloads fast. This is deep, low-level work (performance counters, memory bandwidth, warp-level scheduling) applied to the specific shapes and patterns our models actually use.
We hire kernel engineers because the gap between "this works" and "this is fast on the hardware we have" is enormous, and that gap directly bounds what our researchers can try. You'll close that gap.
WHAT YOU'LL DO
Write and optimize GPU kernels (CUDA, ROCm, Triton, or similar) for training and inference workloads: attention variants, MoE layers, custom activations, communication primitives
Profile real workloads with hardware counters and translate findings into specific kernel-level optimizations
Co-design kernels with the research teams, when the kernel and the algorithm need to change together, you participate in both
Integrate optimized kernels into our training and serving stacks; benchmark before and after; verify the win is real end-to-end
Maintain kernel quality over time as hardware, frameworks, and workloads shift underneath
Spread kernel-level fluency across the team; we want this expertise shared, not siloed
WHAT WE'RE LOOKING FOR
4+ years writing performant GPU kernels (CUDA, ROCm, Triton, or production-grade equivalent)
