LLM Performance Architect (Pre-Silicon, GPU Systems)

NVIDIA Corporation; United States; On-site; USD 152,000 - 241,000; Full time

Benefits offered by this job

Equity

Benefits

Job summary

NVIDIA Corporation is seeking New College Graduates for a Deep Learning Performance Architect role focused on pre-silicon modeling and system-level analysis. You will work on modeling workloads for LLM inference, agentic AI, and other generative AI tasks, collaborating across kernel, hardware, and performance teams.

You will build accurate models, improve projection fidelity, and study how GPU microarchitecture impacts efficiency and scalability, contributing to next-generation NVIDIA platforms.

Qualifications

• BS, MS, or PhD in Computer Science, Computer Engineering, Electrical Engineering, or a related field, or equivalent experience.
• Good programming skills in C++ and Python, with solid software engineering fundamentals.
• Basic understanding of modern deep learning models such as Transformers or LLMs.
• Interest in pre-silicon modeling, GPU architecture, correlation, system-level behavior, or systems optimization.
• Exposure to computer architecture, parallel programming, performance analysis, correlation, system modeling, or related systems topics.
• Familiarity with at least one ML framework or profiling tool such as PyTorch, TensorRT-LLM, Triton, Nsight, vLLM.

Responsibilities

• Build or improve pre-silicon performance and system-level models.
• Develop correlation methodologies to compare projections with reference data.
• Study how GPU microarchitecture and system-level design affect end-to-end performance.
• Identify performance bottlenecks across model execution, kernels, data movement, and system behavior.
• Explore architecture ideas to improve performance and efficiency.
• Collaborate with kernel, hardware, and framework teams to understand end-to-end behavior for LLMs and AI models.
• Use profiling and benchmarking tools to support model development and analysis.
• Present findings and insights through clear analysis and visualization.

Skills

C++

Python

CUDA programming

ML frameworks

Education

BS/MS/PhD in CS/CE/EE

Tools

Nsight

TensorRT-LLM

Triton

vLLM

CUDA toolkit

Job description

NVIDIA Corporation is seeking New College Graduates for a Deep Learning Performance Architect role focused on pre-silicon modeling and system-level analysis. You will work on modeling workloads for LLM inference, agentic AI, and other generative AI tasks, collaborating across kernel, hardware, and performance teams.

You will build accurate models, improve projection fidelity, and study how GPU microarchitecture impacts efficiency and scalability, contributing to next-generation NVIDIA platforms.