Senior DL Compiler Engineer -CUDA Tile

NVIDIA•Austin, CA

About The Position

NVIDIA is seeking software engineers for the CUDA Tile team to work on a new tile-based programming model for their GPUs. This role involves designing and implementing compiler transformations, developing MLIR-based dialects and lowering passes, and optimizing the performance of tile-based kernels across various NVIDIA GPU architectures. The responsibilities include defining public APIs, crafting and implementing compiler and optimization techniques, performance optimization, and general software engineering tasks. NVIDIA GPUs are central to advancements in AI, including generative AI, large language models, and more. This is an opportunity to join a top-tier team and make a significant impact on the deep learning community.

Requirements

Bachelors, Masters or Ph.D. in Computer Science, Computer Engineering or a related field (or equivalent experience).
3+ years of relevant work or research experience in compiler optimization, performance analysis and IR design.
Ability to work independently, define project goals and scope, and lead your own development effort.
Excellent C/C++ programming and software design skills, including debugging, performance analysis, and test design.
Strong interpersonal skills.
Ability to work in a dynamic product-oriented team.

Nice To Haves

Knowledge of CPU and/or GPU architecture.
CUDA or OpenCL programming experience.
Experience with MLIR, LLVM, XLA, TVM.
Experience with deep learning models and algorithms.

Responsibilities

Design and implement compiler transformations.
Develop MLIR-based dialects and lowering passes.
Optimize the performance of tile-based kernels for efficient execution across multiple generations of NVIDIA GPU architectures.
Define public APIs.
Craft and implement compiler and optimization techniques.
Performance optimization.
General software engineering work.