Kernel Engineer, AI Accelerator Performance & Optimization

DensityAI; Mountain View (WY); On-site; USD 260,000 - 320,000; Full time

Benefits offered by this job

Equity grant

Medical/dental/vision

401(k)

PTO

Job summary

DensityAI is seeking a software engineer to write, evaluate, and profile specialized compute kernels that run on a custom AI accelerator. You’ll work with architecture and compiler teams to define the kernel programming model and drive performance profiling that informs silicon design.

You’ll write and optimize tensor operations and data movement patterns, develop profiling tools, and contribute to MLIR/LLVM related components.

Qualifications

• Production-grade C/C++ systems code experience.
• Deep GPU kernel experience with CUDA or equivalent.
• Strong computer architecture knowledge.
• Proven performance profiling and optimization skills.
• Experience mapping tensor operations to hardware.

Responsibilities

• Write and optimize compute kernels for an AI accelerator.
• Develop and maintain kernel profiling infrastructure.
• Define shuffle patterns for ML kernel primitives.
• Drive kernel DSL design and memory management strategies.
• Enable end-to-end kernel execution on the architectural simulator.
• Collaborate on the MLIR/LLVM compiler infrastructure and kernel validation.
• Create onboarding documentation and kernel writing guides.

Skills

C/C++

CUDA

Computer architecture

Profiling

Tensor operations

Python

Tools

CUTLASS

MLIR

LLVM

Job description

DensityAI is seeking a software engineer to write, evaluate, and profile specialized compute kernels that run on a custom AI accelerator. You’ll work with architecture and compiler teams to define the kernel programming model and drive performance profiling that informs silicon design.

You’ll write and optimize tensor operations and data movement patterns, develop profiling tools, and contribute to MLIR/LLVM related components.