Senior Backend GPU Compiler Engineer — LLVM/JIT

NVIDIA; Redmond (WA); On-site; USD 152,000 - 242,000; Full time

Job summary

NVIDIA is seeking a Senior Backend Compiler Engineer to advance LLVM-based backend generation for NVIDIA GPUs in the Graphics/Compute stack. You will design new compiler passes, optimize for high performance and low latency, and collaborate with global teams to improve tooling and architecture support.

Ideal candidates have 4+ years in compiler development, strong LLVM code generation expertise, and hands-on C++ skills.

Qualifications

• BS or higher in Computer Science/Engineering or equivalent experience
• 4+ years of compiler development, incl. 3+ years in LLVM with focus on code generation
• Expertise in LLVM code generation: instruction scheduling, software pipelining, register allocation, GlobalISel, TableGen, LLVM IR, and MIR
• Proficient in compiler design and performance analysis tools
• Excellent hands-on C++ programming skills with industry standard C++ compilers
• Strong software engineering fundamentals, robust and maintainable solutions
• Good communication and documentation skills and self-motivated

Responsibilities

• Guide design and implementation of a new LLVM-based backend JIT compiler for NVIDIA GPUs
• Design and develop new compiler passes and optimizations for high performance and low compilation time
• Collaborate with global teams to oversee improvements and resolutions
• Apply and adapt compiler technologies for premier, robust tools
• Influence and work on new GPU architectures

Skills

LLVM code generation

C++ programming

Compiler design

Performance analysis

Communication & documentation

Education

Bachelor's degree in CS/Engineering

Masters or PhD preferred

Tools

GlobalISel

TableGen

LLVM IR

Machine IR (MIR)

Performance profiling tools

Job description

NVIDIA is seeking a Senior Backend Compiler Engineer to advance LLVM-based backend generation for NVIDIA GPUs in the Graphics/Compute stack. You will design new compiler passes, optimize for high performance and low latency, and collaborate with global teams to improve tooling and architecture support.

Ideal candidates have 4+ years in compiler development, strong LLVM code generation expertise, and hands-on C++ skills.