NVIDIA is at the forefront of the generative AI revolution! The Algorithmic Model Optimization Team specifically focuses on optimizing generative AI models such as large language models (LLM) and diffusion models for maximal inference efficiency using techniques ranging from neural architecture search and pruning to sparsity, quantization, and automated deployment strategies. Our work includes conducting applied research to improve model efficiency as well as developing an innovative software platform (TRT Model Optimizer). Our software is used both internally across NVIDIA and externally by research and engineering teams alike developing best-in-class AI models.

We are now looking for a Senior Deep Learning Software Engineer to develop and scale up our automated inference and deployment solution. As part of the team, you will be instrumental in pushing the limits of inference efficiency and large-scale, automated deployment. Your work will touch upon fundamental aspects of a typical machine learning stack including working in high-level frameworks like PyTorch and HuggingFace to developing and improving high-performance kernel implementations in CUDA, TRT-LLM, and Triton. This is an exceptional opportunity for passionate software engineers straddling the boundaries of research and engineering, with a strong background in both machine learning fundamentals and software architecture & engineering.

• Train, develop, and deploy state-of-the generative AI models like LLMs and diffusion models using NVIDIA's AI software stack.
• Leverage and build upon the torch 2.0 ecosystem (TorchDynamo, torch.export, torch.compile, etc...) to analyze and extract standardized model graph representation from arbitrary torch models for our automated deployment solution.
• Develop high-performance optimization techniques for inference, such as automated model sharding techniques (e.g. tensor parallelism, sequence parallelism), efficient attention kernels with kv-caching, and more.
• Collaborate with teams across NVIDIA to use performant kernel implementations within our automated deployment solution.
• Analyze and profile GPU kernel-level performance to identify hardware and software optimization opportunities.
• Continuously innovate on the inference performance to ensure NVIDIA's inference software solutions (TRT, TRT-LLM, TRT Model Optimizer) can maintain and increase its leadership in the market.
• Play a pivotal role in architecting and designing a modular and scalable software platform to provide an excellent user experience with broad model support and optimization techniques to increase adoption.

• Masters, PhD, or equivalent experience in Computer Science, AI, Applied Math, or related field.
• 5+ years of relevant work or research experience in Deep Learning.
• Excellent software design skills, including debugging, performance analysis, and test design.
• Strong proficiency in Python, PyTorch, and related ML tools (e.g. HuggingFace).
• Strong algorithms and programming fundamentals.
• Good written and verbal communication skills and the ability to work independently and collaboratively in a fast-paced environment.

• Contributions to PyTorch, JAX, or other Machine Learning Frameworks.
• Knowledge of GPU architecture and compilation stack, and capability of understanding and debugging end-to-end performance.
• Familiarity with NVIDIA's deep learning SDKs such as TensorRT.
• Prior experience in writing high-performance GPU kernels for machine learning workloads in frameworks such as CUDA, CUTLASS, or Triton.