Efficient Large-Scale LLM Inference Research Intern – Huawei Switzerland Research Centre, Lausanne

About Huawei
Huawei is a leading global ICT (information and communications technology) solutions provider. We’re driven by innovation, open collaboration, and strong operations. Our ICT solutions, products, and services are used in over 170 countries and regions, serving more than one-third of the world’s population. With 180,000 employees, Huawei is committed to enabling the future information society and building a Better Connected World.

Huawei’s Research Centre in Lausanne focuses on advanced technical research, architecture evolution design, and strategic technical planning in computer architecture.

Responsibilities

• Speed up or improve performance of inference LLM systems at the system level.
• Develop highly efficient modules for attention and other core LLM components in CUDA or C++.
• Collaborate with leaders in the field to design new core kernels to speed up inference.
• Perform profiling and analysis of distributed LLMs.
• Work with state-of-the-art LLM models (e.g, DeepSeek, MoE, ChatGPT) and inference frameworks (e.g., vLLM).

Requirements

• Currently enrolled in a Master’s or PhD program in CS, Computer Engineering, or related fields
• Experience in C/C++, and ideally CUDA
• Knowledge of core ML models (e.g., transformers, attention, ...)
• Knowledge of distributed systems (e.g., shared memory, MPI, , ...) and/or parallel programming (e.g., OpenMP, pThreads) is a plus
• Ideally, experience with LLMs, HPC
• Familiarity with AI frameworks (PyTorch, TensorFlow, JAX) is a plus
• Excellent English communication
• Good analytical skills, willing to learn, good coding expertise, research-oriented mindset

By applying to this position, you agree with our PRIVACY STATEMENT. You can read in full our privacy policy at https://www.huawei.com/en/privacy-policy.