LLM Algorithmic Optimization Engineer

NIO Inc. is a pioneer and a leading company in the global smart electric vehicle market.

Founded in November 2014, NIO aspires to shape a sustainable and brighter future with the mission of “Blue Sky Coming”.

NIO envisions itself as a user enterprise where innovative technology meets experience excellence.

NIO designs, develops, manufactures and sells smart electric vehicles, driving innovations in next-generation core technologies.

NIO distinguishes itself through continuous technological breakthroughs and innovations, exceptional products and services, and a community for shared growth.

NIO provides premium smart electric vehicles under the NIO brand, premium smart electric vehicles for families through the ONVO brand, and small smart high-end electric cars with the FIREFLY brand.

• Conduct research and apply cutting-edge technologies to optimize Large Language Models (LLMs) and multimodal models, exploration and implementation of the core algorithmic optimization on heterogeneous architectures, for highly efficient LLM inference as well as deployment across distributed and heterogeneous hardware environments.
• Focus on model optimization from a systems perspective, ensuring efficient deployment in the vehicle’s digital cockpit and advanced driving (AD) domain.
• Collaborate with cross-functional teams to ensure the integration of optimized models into real-world automotive applications.
• Contribute to the entire pipeline from research, development, and testing, through to deployment on hardware, including GPUs and other distributed systems.
• Currently pursuing or completed a PhD or Master’s degree in Computer Science, Computer Engineering, Applied Mathematics, Communications, Electronics, or a related field with relevant research projects and publications.
• Strong understanding of GPU/NPU architecture and optimization techniques to identify and address bottlenecks.
• Proficient in LLM and VLM architectures and algorithms, familiar with transformer based NLP / Audio / CV algorithms and technologies.
• Proficiency in Python and experience with AI-related training and inference tools such as PyTorch.
• Proficiency in C/C++ programming, familiar with at least one commonly used LLM inference engines.
• Hands-on experience with model-serving frameworks such as Open Neural Network Exchange (ONNX).
• Familiarity with debugging code in distributed computing environments.Experience in LLM inference optimization on resource constrained edge devices is a plus.