LLM Algorithm Engineer

About the role

EveryMatrix is seeking an LLM Algorithm Engineer to join its EM Labs team in Changsha, China. You will work on cutting‑edge large language model deployment, optimization, and productisation within a fast‑growing iGaming SaaS environment.

Key responsibilities

• Design and implement distributed deployment solutions on NVIDIA hardware (NVLink/NVSwitch) and lead framework selection for vLLM, TensorRT‑LLM, SGLang.
• Build and manage multi‑modal GPU clusters, optimizing KV‑Cache storage and loading for long‑context inference.
• Perform model optimization, fine‑tuning with LoRA/QLoRA, and integrate Retrieval‑Augmented Generation (RAG) techniques.
• Develop enterprise‑grade internal toolchains such as code‑assist, code‑review, and private knowledge systems.
• Design external customer‑facing platforms, including smart customer‑service solutions and multi‑agent collaborative assistants.

Required profile

• Master’s degree or higher in Computer Science, Artificial Intelligence or a related field.
• At least 3 years of professional experience on NLP/LLM projects.
• Strong understanding of Transformer architectures and attention‑mechanism optimization.
• Familiarity with CUDA programming and NVIDIA topology design; experience with NVIDIA kernel optimization is a plus.

Required skills

• PyTorch
• TensorFlow
• Transformer architecture
• CUDA
• NVLink / NVSwitch
• vLLM, TensorRT‑LLM, SGLang
• LangChain, AutoGen
• LoRA, QLoRA, PEFT
• Prompt Engineering
• Retrieval‑Augmented Generation (RAG)
• GPU cluster management