Senior Performance Architect, Nemotron

What you'll do

• Develop high-fidelity performance models to evaluate emerging AI techniques and hardware optimizations.
• Prioritize features based on detailed performance modeling to guide future software and hardware roadmaps.
• Model end-to-end performance impacts of new GenAI workflows to predict datacenter needs.
• Keep abreast of the latest DL research and collaborate with diverse teams for co-design decisions.
• Define metrics, design experiments, and visualize large datasets to identify resource bottlenecks in AI workloads.

What we're looking for

• Master's degree or equivalent experience in Computer Science or related fields.
• Expertise in computer architecture, roofline modeling, queuing theory, and statistical performance analysis.
• Strong background in ML fundamentals, model parallelism, and inference serving techniques.
• Proficiency in Python and C++ for simulator design and data analysis.
• 3+ years of experience in system evaluation of AI/ML workloads or performance optimization.
• Experience with deep learning frameworks like PyTorch, TRT-LLM, VLLM, SGLang.
• Comfortable defining metrics, designing experiments, and visualizing large datasets to identify bottlenecks.