Software Engineer, DL Kernel Libraries - New College Grad 2027
NVIDIAJob Title
Software Engineer, DL Kernel Libraries - New College Grad 2027
Role Summary
Join the DL Kernel Library team to develop deep learning inference and training systems software such as cuDNN and FlashInfer. The team designs library abstractions, kernel implementations, and code generators to accelerate large language models and other AI workloads on NVIDIA GPUs.
This is an early-career, full-time engineering role working closely with framework, library, kernel, and GPU architecture teams to deliver high-performance runtime and compiler technologies.
Experience Level
Entry-level - new college graduate (class of 2027) position; suitable for engineers recently completing BS/MS/PhD or with equivalent experience.
Responsibilities
The role focuses on building and optimizing kernels, runtimes, and compilers for AI workloads. Typical responsibilities include:
- Design, implement, and optimize GPU kernels for inference and training workloads.
- Develop extensible abstractions for serving engines and training frameworks.
- Build just-in-time, domain-specific compilers and runtime components.
- Collaborate across teams (frameworks, libraries, GPU architecture) to integrate and validate solutions.
- Contribute to and interact with open source projects used by the stack (examples: FlashInfer, cuDNN, vLLM, SGLang, PyTorch, JAX, XLA).
Requirements
Key must-have skills and experience for successful candidates; follow-on items are desirable but optional.
- Must-have: Strong experience using or developing deep learning frameworks (PyTorch, JAX, TensorFlow, ONNX) or inference engines and runtimes (vLLM, SGLang, MLC).
- Must-have: Strong programming skills in Python and C/C++.
- Must-have: Ability to work in a collaborative engineering environment across software and hardware teams.
- Nice-to-have: Experience with domain-specific compilers or library solutions for LLM inference and training (e.g., FlashInfer, Flash Attention).
- Nice-to-have: Familiarity with machine learning compilers (Apache TVM, MLIR) and GPU kernel performance optimization (CUDA C/C++, CUTLASS DSL, cuTile, Triton, etc.).
- Nice-to-have: Open source project contributions or ownership.
Education Requirements
BS, MS, or PhD in Computer Science, Electrical Engineering, or a related technical field - or equivalent practical experience.
About the Company
Company: NVIDIA
Headquarters: Santa Clara, California, USA
NVIDIA is a global leader in accelerated computing, renowned for its innovative solutions in AI and digital twins that transform diverse industries. The company specializes in networking technologies, providing end-to-end InfiniBand and Ethernet solutions for servers and storage that optimize performance and scalability. NVIDIA serves sectors such as high-performance computing, enterprise data centers, and cloud computing, constantly reinventing its products and services to stay ahead in the market.
