Startups · AI

Compiler Engineer-MLIR

Density · Mountain View, CA · On-site

← All jobs
About Density

Density shows you how people use your space and helps you get more out of every square foot. Backed by Kleiner Perkins and South Park Commons.

About the role

Own MLIR dialect design and lowering passes for our AI accelerator — defining the high-level tensor IR, async / streaming semantics, and sharded-tensor types that bridge ML frameworks to silicon. Work with chip-design and software teams driving DensityAI's AI accelerator program from first silicon through scale-out.

What they're looking for

  • Exceptional abilities in MLIR dialect design, lowering pass authoring, and rewrite patterns
  • 5+ years compiler engineering experience, with hands-on MLIR contributions or equivalent IR-design experience
  • Deep understanding of tensor compilation, distributed / sharded execution, and async / streaming dataflow models
  • Strong C++ fluency and experience integrating with ML frameworks (PyTorch, JAX, ONNX, TensorFlow, or equivalent)
  • (Optional) LLVM backend experience, GPU compiler experience (Triton, IREE, XLA, or equivalent), or open-source MLIR / LLVM contributions
More about this role

Own MLIR dialect design and lowering passes for our AI accelerator — defining the high-level tensor IR, async / streaming semantics, and sharded-tensor types that bridge ML frameworks to silicon. Work with chip-design and software teams driving DensityAI's AI accelerator program from first silicon through scale-out.

  • Own MLIR dialect design and lowering passes for our AI accelerator — defining the high-level tensor IR, async / streaming semantics, and sharded-tensor types that bridge ML frameworks to silicon
  • Use and develop AI-assisted tool flows to accelerate compiler development
  • Exceptional abilities in MLIR dialect design, lowering pass authoring, and rewrite patterns
  • 5+ years compiler engineering experience, with hands-on MLIR contributions or equivalent IR-design experience
  • Deep understanding of tensor compilation, distributed / sharded execution, and async / streaming dataflow models
  • Strong C++ fluency and experience integrating with ML frameworks (PyTorch, JAX, ONNX, TensorFlow, or equivalent)
  • (Optional) LLVM backend experience, GPU compiler experience (Triton, IREE, XLA, or equivalent), or open-source MLIR / LLVM contributions

Final offers depend on level,...

Read the full posting on Density's site ↗

Software

Build your edge while you search

Free tools for founders and investors, plus VC Unfiltered, our take on startups, venture and the people who build them.