Deep Learning Performance Architect - New College Grad 2027 - CA, Santa Clara, United States
$124,000 - $241,500 USD yearly
Originally posted 27 October 2026 by the employer.
About the role
As a Deep Learning Performance Architect, you will join the GPU compute architecture and performance engineering team focused on pre-silicon modeling, data-supported architecture exploration, and GPU microarchitecture improvement for LLM inference, agentic AI, and generative AI workloads. You will build system-level modeling solutions to analyze efficiency, scalability, and performance bottlenecks across NVIDIA platforms.
What you'll do
- Build and improve pre-silicon performance and system-level models to analyze workload behavior, architecture tradeoffs, and projection accuracy. - Develop correlation methodologies to compare model projections with lower-level analyses or reference data to enhance model fidelity. - Study how GPU microarchitecture and system-level design choices affect end-to-end performance, efficiency, scalability, and projection fidelity. - Identify performance bottlenecks across model execution, kernels, data movement, and system behavior. - Explore and evaluate architecture ideas, collaborating with kernel, hardware, and framework teams across LLMs and agentic AI systems. - Use profiling and benchmarking tools to support model development, analysis, and correlation, presenting findings through clear visualization.
What you'll need
- BS, MS, or PhD in Computer Science, Computer Engineering, Electrical Engineering, or a related field, or equivalent experience. - Programming skills in C++ and Python, with solid software engineering fundamentals. - Basic understanding of modern deep learning models such as Transformers or LLMs. - Exposure to computer architecture, parallel programming, performance analysis, correlation, or system modeling through coursework, research, internships, or projects. - Familiarity with at least one ML framework, modeling environment, profiling tool, or software stack focused on efficiency, such as PyTorch, TensorRT-LLM, Triton, Nsight, or vLLM. - Ability to reason about performance tradeoffs and debug technical issues.
Skills: pre-silicon performance modeling, GPU microarchitecture, C++, Python, LLMs, deep learning
This role has been open 0 days — well below the 202-day median for SoC/Chip Architecture roles.
SoC/Chip Architecture · AI ML Hardware
|
Open roles in category
31
|
Median days open
202 d
|
Median salary
$248k
|
See the full market breakdown ▾Category comparison, skills in demand, and who else is hiring
| Metric | NVIDIA | All employers we track in this specialty (31 roles · 11 employers) |
|---|---|---|
| Open roles in this specialty | 8 | 31 |
| Open roles in the wider Digital Design & Verification family | 234 | 3383 · 116 employers |
| Median days open | 196 d | 202 d (−6 d vs this employer) |
| Median salary (USD postings) | — | $248k |
Skills observed across this category: pre-silicon performance modeling, GPU microarchitecture, C++, Python, LLMs, deep learning
Who's hiring in this category
- Qualcomm · 10 open roles · median 202 d
- NVIDIA (this employer) · 7 open roles · median 198 d
- SiFive · 3 open roles · median 199 d
- AMD · 2 open roles · median 199 d
- Infineon Technologies · 2 open roles · median 196 d
- Intel Corporation · 2 open roles · median 4 d
How we counted: 31 open SoC/Chip Architecture (AI ML Hardware) roles from 11 employers tracked in the SemiconductorJobs index, counted 28 Oct 2026. Specialty figures count only roles carrying this exact specialty label, so an employer's related work in neighbouring specialties is not included there — it is counted in the wider Digital Design & Verification family row. Figures refresh nightly.