Senior Deep Learning Software Engineer, DLSim

NVIDIA•Austin, CA

23h

About The Position

The DL Performance Modeling Team's core mission is to deliver full stack simulation infrastructure for deep learning applications across a spectrum of GPUs. We actively collaborate with architecture, software, product, and research teams to shape and refine the strategic roadmap of DL hardware and software. We are now looking for a Deep Learning Software Engineer to help develop simulation infrastructure. The software rapidly assesses new AI-accelerating GPU hardware and software advancements.

Requirements

A Masters (or equivalent experience) in Computer Science, Computer Engineering, or a related STEM field; PhD preferred.
3+ years of relevant experience in compiler optimization, architectural simulation, or related areas.
Strong hands-on experience with MLIR and compiler infrastructure.
Excellent C/C++ and Python programming skills, including software design, debugging, performance analysis, and test development.
Strong communication and collaboration skills, with the ability to thrive in a fast-paced, multi-functional, outcome-focused environment.

Nice To Haves

Experience designing and building compiler frameworks or intermediate representations from the ground up.
Deep understanding of LLM inference workloads and their implications for computer architecture.
Hands-on experience implementing and optimizing complex AI workloads on CPUs, GPUs, or custom accelerators.

Responsibilities

Develop simulation backends that enable fast, scalable evaluation of AI workloads across NVIDIA compiler stacks.
Improve deep learning compiler kernel code generation and computational graph optimization using analysis based on modeled scenarios and performance insight.
Advance the modeling and optimization of datacenter-scale AI workloads and deployment scenarios.
Partner with architects and software teams to evaluate future GPU features and guide silicon and system-level design decisions.