Deep Learning Performance Architect - New College Grad 2027
Core
Build pre-silicon performance models and system-level analysis to improve efficiency and scalability of LLMs and agentic AI on NVIDIA GPU platforms.
Role type
Junior IC deep learning performance architect (pre-silicon modeling)
Builds
Pre-silicon performance models, system-level correlation solutions, and architectural tradeoff analyses for generative AI workloads
Domain
Semiconductor hardware + Generative AI (LLMs, Agentic AI)
Deliverable
production ML models
Required skills
C++, Python, Computer Architecture, Parallel Programming, Performance Analysis, System Modeling, Profiling Tools, ML Frameworks (PyTorch, TensorRT-LLM, Triton, Nsight, vLLM), GPU Microarchitecture
Preferred skills
CUDA Programming, GPU Performance Analysis, Research Projects, Open-source Contributions
Technologies
CUDA, 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 GPU microarchitecture effects on performance and efficiency; Identify performance bottlenecks across kernels and data movement; Explore architecture ideas to improve model fidelity; Collaborate with kernel and hardware teams; Use profiling tools to support analysis; Present findings through visualization
Seniority
Entry-level (New College Grad 2027)
