Samsung
Staff Engineer, Compiler
$163K–$253K
San Jose, California, United States · Staff · Full-time · Posted Jun 2, 2026
- Experience
- Bachelor’s with 10+ years, or Master’s with 8+ years, or PhD with 5+ years of industry experience; additionally, 3–5+ years in at least one listed compiler technology or equivalent. More experienced candidates may be considered at relevant levels.
- Education
- Bachelor’s, Master’s, or PhD.
About the job
Develop compiler infrastructure and kernel programming tools for Samsung hardware, optimizing AI/ML workloads across its memory hierarchy and distributed systems.
What you'll do
- Adapt torch.compile and lower compiler IR and kernel DSLs to Samsung hardware, ISA, memory hierarchy, and collective primitives.
- Design placement, scheduling, parallelism-aware lowering, fusion, tiling, and memory-planning passes.
- Contribute compiler and DSL improvements to open-source projects, including through upstream review processes.
What they ask for
- 3–5+ years of industry experience with at least one of Triton, Helion, MLIR, XLA, TVM, Inductor, IREE, CUTLASS, or a proprietary equivalent.
- Experience designing a kernel DSL or its IR, or making substantial language-level changes to one.
- Experience with MLIR, including writing dialects, passes, or backend integration.
- Experience building PyTorch backends for non-CUDA accelerators.
Nice to have
- Kernel autotuning, performance modeling, or cost-based compilation experience; background in HPC, distributed systems, or NUMA-aware programming.
- Open-source contributions to PyTorch, Triton, Helion, LLVM/MLIR, or similar projects.
Pay
$163,000 — $253,000 USD base salary
Tools
torch.compile, Inductor, Triton, Helion, MLIR, XLA, TVM, IREE
AI focus
Build compiler and kernel infrastructure to improve the performance, scalability, and energy efficiency of AI/ML workloads.
Model-written summary · the original posting is authoritative