Member of Technical Staff, GPU Kernels
$285k - $315k • Remote • San Francisco • FullTime
Posted 1mo ago
Job Location
San Francisco
Tech Stack
Remote Work Policy
Fully remote
Employment Type
FullTime
Categories
Machine Learning Engineer
About the job
SF Tensor is building the future of high-performance compute for AI by rethinking and rebuilding the entire stack, from hardware to cloud. We aim to eliminate the friction that slows down AI development, making compute faster, cheaper, and more accessible. Our Kernel Optimizer automatically finds the fastest code for any hardware, and our Model Foundry manages workloads across clouds and chips. We are backed by leading investors and are seeking individuals passionate about advancing AI through compute innovation.
Responsibilities
- Write and hand-optimize GPU kernels for pre-training, post-training, and inference across various hardware platforms (NVIDIA, AMD, TPU, Trainium).
- Translate hand-optimized kernel learnings into machine-searchable structures for the compiler.
- Profile GPU performance at the microarchitectural level, analyzing metrics like SM/CU utilization, warp stalls, and register pressure.
- Debug performance issues down to clock behavior, thermal throttling, and driver paths.
- Work at the ISA level (SASS, cubins) below PTX to emit optimized schedules.
- Build performance models, microbenchmarks, and tooling to predict kernel behavior.
- Collaborate with the formal correctness team to ensure aggressive kernels are shipped with formal proofs.
Requirements
- Proven track record of hand-writing GPU kernels that match or exceed vendor library performance.
- Comfortable reading and understanding PTX, SASS, GCN/CDNA ISA, or equivalent machine-level assembly.
- Proficiency with low-level GPU profiling tools (Nsight Compute, Nsight Systems, rocprof, omniperf, or equivalents).
- Solid systems programming skills in C++ and CUDA or ROCm/HIP.
- Working understanding of how high-level ML operations map to hardware and the impact of framework layers.