Senior AI Compiler Engineer, MLIR
Core
Build an MLIR-based AI compiler for NVIDIA's inference engine, focusing on performance, fast builds, low memory use, and AOT/JIT usability across data center and edge.
Role type
Senior IC AI compiler engineer (MLIR)
Builds
MLIR-based graph representations, optimizations, and kernel generation for neural networks
Domain
AI computing, compiler technology, GPU architecture
Deliverable
production ML models
Required skills
MLIR, XLA, LLVM, C/C++, Python, performance analysis, compiler optimizations, software design
Preferred skills
deep learning models/algorithms/frameworks (PyTorch, JAX), GPU kernel generation, CUDA/OpenCL, mentoring
Technologies
MLIR, XLA, LLVM, CUDA, OpenCL, PyTorch, JAX
Responsibilities
Develop MLIR-based graph representations and optimizations for future GPU architectures; Partner with framework and hardware teams to enable new model patterns and GPU features; Define APIs and MLIR dialects; Conduct performance optimizations and analysis; Implement compiler optimizations and kernel generation for neural networks
Seniority
Senior, hands-on IC