Software Engineer, DL Kernel Libraries - New College Grad 2027

at Nvidia

USD 108,000-195,500 per year

JUNIOR

✅ On-site

Tech Stack
Tag name is followed by "@" symbol and proficiency level value. About proficiency levels:

1-2 — basic awareness. Minimal hands-on experience, and a rudimentary understanding of the technology's purpose;

3-6 — daily use. Comfortable and regular usage, capable of handling common tasks and challenges related to the technology;

7-9 — you are an expert, you can teach others, you know all the pitfalls and tricks;

10 — exceptional knowledge, comprehensive understanding, and adeptness in all aspects of the technology, including advanced problem-solving. Think twice before claiming or demanding such level.

AI CUDA @ 6 Deep Learning @ 6 GPU @ 6 JAX @ 6 LLM @ 3 Machine Learning @ 3 Performance Optimization @ 6 PyTorch @ 6 Python @ 6 SGLang @ 6 TensorFlow @ 6 vLLM @ 6

Details

NVIDIA is looking for software engineers to develop deep learning inference and training systems software. The DL Kernel Library team builds AI systems technologies, including cuDNN and FlashInfer, as well as libraries, code generators, and GPU kernel technologies for NVIDIA hardware architectures. The work includes developing abstractions, efficient kernel implementations, and kernel code generators to accelerate large language models, agents, and other AI workloads.

Responsibilities

Innovate and develop AI systems technologies for efficient inference and training.
Design, implement, and optimize kernels for high-impact AI workloads.
Design and implement extensible abstractions for serving engines and training frameworks.
Build efficient just-in-time domain-specific compilers and runtimes.
Collaborate with engineers across deep learning frameworks, libraries, kernels, and GPU architecture teams.
Contribute to open-source communities and projects including FlashInfer, cuDNN, vLLM, SGLang, PyTorch, JAX, and XLA.

Requirements

Bachelor's, master's, or PhD degree in Computer Science, Electrical Engineering, or a related field, or equivalent experience.
Strong experience developing or using deep learning frameworks such as PyTorch, JAX, TensorFlow, or ONNX, or inference engines and runtimes such as vLLM, SGLang, and MLC.
Strong Python and C/C++ programming skills.

Preferred Qualifications

Background in domain-specific compiler and library solutions for LLM inference and training, such as FlashInfer and FlashAttention.
Expertise in training frameworks and inference engines.
Expertise in machine learning compilers such as Apache TVM and MLIR.
Strong experience in GPU kernel development and performance optimization, especially with CUDA C/C++, CUTLASS DSL, cuTile, Triton, or similar technologies.
Open-source project ownership or contributions.